The Sonix RP is an easy to use ultrasound imaging device that is capable of acquiring high quality images. In addition to supporting the acquisition of multiple data types such as RF, elastography, and color doppler data, the machine is an open ended system providing users with full control over imaging parameters through an investigational research interface. Since the Sonix RP is PC based and it supports open-source software development toolkits, programs can be developed and executed directly onto the device, thus eliminating the need for extra hardware that is often required for data collection and processing. Due to these advantages, many universities and research institutes have successfully used the Sonix RP to test and implement their customized solutions for different applications.

Dual-wavelength imaging is used in several scientific and practical applications. One of the most common applications is dual-wavelength thermography which has many advantages over single wavelength thermal imaging. Optical imagesplitters can be used to turn any imaging equipment into a dual-wavelength imaging system. In this paper, a new design of an image-splitting optic, for use in dual-wavelength imaging, is presented. The new design evades the limitations encountered with the basic image-splitter design where images can be captured at higher resolutions and frame rates. The new design also facilitates the adjustment of the image magnification. With very minor changes in the optical components, the image-splitter can be used in different thermal imaging techniques such as Infrared (IR) imaging and Laser Induced Fluorescence (LIF) imaging or any other technique that utilizes dual-wavelength imaging. Furthermore, with some modifications in the optical path, the image splitter can be used for imaging

The variable focus telescope is utilized now in the laser transmitter system design. It changes the telescope's magnifying power in order to adjust the exit beam through moving the inner focus lens. This system has complicated structure and high machining expense. This paper investigates the focusing character of Gaussian beam through misadjust telescope and presents a new method for lidar transmission system design. The laser beam divergence angle and the radius of exit beam are changed through moving the distance between the back focus of object lens and the front focus of ocular. This design can provide a convenient method for calculating the focusing parameters. The restriction of assembly dimension and the effect of fitting and adjusting error which should be considered in engineering application are studied, and then bring a method to choose the right parameters of focusing optic system by the focusing requirement.

This paper presents a large set of spectral and directional signatures of the polarized reflectance acquired over various plant canopies in different atmospheric conditions. An instrument has been developed for measuring the BPDF (Bidirectional Polarization Distribution Function) of plant canopies in the field. Polarized multi-wavelength analytical physically-based model was developed. For the analysis of polarization measurements studied, it is found that although spectral variations in the polarized reflectance are observed, the ratio of the two wavelength polarized reflectance is stable. The ratio is related to atmospheric aerosol optical depth. Our results also suggest that using the correlation between the polarized reflectance of the short wave infrared band (SWIR) with those in the visible rang can eliminate the effect which caused by the plant canopies geometric structure. On the other hand, since the model accurately predict the polarized reflectance relations between the short wave infrared bands and the visible rang, they can be used to discriminate the aerosol contribution from the surface of the plant canopies cover in the retrieval procedure.

A liquid crystal lens with wide spectrum electrical tunable focal length is proposed. The top substrate coated indium tin oxide film with a hole of 2.0mm diameter is manipulated by techniques mainly consists of conventional ultraviolet lithography and hydrochloric acid etching. The bottom substrate glass coated ITO acts as another electrode. The nonuniform electric field appeared in the LC layer produces a gradient distribution of refractive index. The LC cell shows an optical lens property. The focal length of the LC lens is function of applied voltage, covering approximately 20mm- 480mm. In the focal length range, the LC lens maintains high optical quality. The transmissivity of the LC lens is above 80% from 500nm to 1100nm.

Characteristics of Geosynchronous Synthetic Aperture Radar (GEO SAR), including ground velocity, integration time, azimuth resolution and other factors, are investigated. And the relationship between them and orbital elements is analyzed. Then a subaperture mode based on yawing steering which is more suitable for GEO SAR is proposed. By applying this mode reduction of range migration is obvious, and system parameters of L-band GEO SAR are designed, including the antenna size, the pulse width, the peak transmitted power, and pulse repetition frequency. Finally a modified Range Doppler (RD) algorithm based on quintic polynomial and a Chirp Scaling (CS) algorithm based on spectrum mosaic are advanced. Point target simulation is implemented and the azimuth resolution is assessed according to the rule of being able to distinguish adjacent targets. It is demonstrated that the proposed image formation algorithms can provide a spatial resolution better than 15m.

Terahertz radiation is a special electromagnetic spectrum range sandwiched between the infrared ray and microwave. With developing of super fast and super power laser technology, this special range has been explored extensively. Among all terahertz technologies, terahertz imaging may be the first one can be used in the practical applications. In this presentation, a novel terahertz multi-spectral focal plane imaging method is proposed. Using a CCD camera instead of the single detector, this method can capture the time domain waveform of an object quickly. Using the fast Fourier transform, both the amplitude and phase of the THz spectrum can be achieved. Based on this information, the layer structure inside the object can be presented. The multi-spectral phase imaging technology has also been employed to get high signal noise ration.

Passive millimeter-wave (PMMW) imaging offers advantages over visible and IR imaging in having better all weather performance. However the PMMW imaging sensors are state-of-the-art to date, sometimes it is required to predict and evaluate the performance of a PMMW sensor under a variety of weather, terrain and sensor operational conditions. The PMMW scene simulation is an efficient way. This paper proposes a framework of the PMMW simulation for ground scenes. Commercial scene modeling software, Multigen and Vega, are used to generate the multi-viewpoint and multi-scale description for natural ground scenes with visible images. The background and objects in the scene are classified based on perceptive color clusters and mapped with different materials. Further, the radiometric temperature images of the scene are calculated according to millimeter wave phenomenology: atmospheric propagation and emission including sky temperature, weather conditions, and physical temperature. Finally, the simulated output PMMW images are generated by applying the sensor characteristics such as the aperture size, data sample scheme and system noise. Tentative results show the simulation framework can provide reasonable scene's PMMW image with high fidelity.

Soil moisture content is one of the most important factors in soil business. The basic of detecting soil moisture content using remote sensing technology is to analyze the relationship between soil moisture content and emissivity. In this paper, based on the analysis of spectrum collection and processing by a portable spectrometer, a set of measure schemes were first established which can accurately measure the reflectivity and emissivity of soil spectrum with different moisture content in near-infrared and thermal infrared bands. Then we selected different bare soil areas as the areas for survey, and studied the relationship of different moisture content and the spectrum curve in the soil both of the same kind and of different kind (like the soil whose structure has been modified caused by the change of organic matter contents or soil particle size). Finally, we emphasized on the quantitative relationship between soil reflectivity & emissivity and soil moisture content using the test data, and establish a model depicting the quantitative relationship above in near-infrared and thermal infrared bands.

Infrared image simulation plays an important role in many fields. This paper presents a method for dynamic infrared image simulation under various natural conditions. Firstly, we analyze the infrared imaging rationale, considering various affected factors of temperature, radiance filed and atmospheric attenuation based on the heat transfer equation. Secondly, we present a reasonable simulation procedure based on the Vega Application Interface (API). Especially, a hybrid simulation idea is put forward and an improved strategy for different objects is carried out. The performance of this approach is evaluated by simulating two kinds of targets. Results show that the method proposed is valid compared with the real images under the same condition. In addition, this approach has relatively low calculation complexity but high fidelity, and satisfies the requirement of real time.

X wave has a large depth of field and may have important application in ultrasonic imaging to provide high frame rate (HFR). However, the HFR system suffers from lower spatial resolution. In this paper, a study of nonlinear imaging with X wave is presented to improve the resolution. A theoretical description of realizable nonlinear X wave is reported. The nonlinear field is simulated by solving the KZK nonlinear wave equation with a time-domain difference method. The results show that the second harmonic field of X wave has narrower mainlobe and lower sidelobes than the fundamental field. In order to evaluate the imaging effect with X wave, an imaging model involving numerical calculation of the KZK equation, Rayleigh-Sommerfeld integral, band-pass filtering and envelope detection is constructed to obtain 2D fundamental and second harmonic images of scatters in tissue-like medium. The results indicate that if X wave is used, the harmonic image has higher spatial resolution throughout the entire imaging region than the fundamental image, but higher sidelobes occur as compared to conventional focus imaging. A HFR imaging method with higher spatial resolution is thus feasible provided an apodization method is used to suppress sidelobes.

Distortion compensation of High-resolution range profile plays an important role in ISAR imaging of high-speed space target. Since the baseband echo of such target is a multi-component linear frequency modulated(LFM) signal with identical chirp rates. Since the peaks of its ambiguity function lie on a set of parallel lines, whose slope is relative to the chirp rate, an estimator based on the fractional autocorrelation is provided to estimate the target's radial velocity. Furthermore, the particle swarm optimization(PSO) algorithm is also applied to speed up the progress of parameter estimation,. Finally, the velocity compensation is accomplished with the estimated velocity. Simulations show that the proposed method can compensate the range profile distortion of space target effectively.

A W-band integrated multi-channel receiver for passive Millimeter-wave (PMMW) imaging application is developed. The construction and operating principle of multi-channel receiver module is described. For each channel, heterodyne receiver is realized in order to reach high stability, which is very important to PMMW system. High integration of the multi-channel receiver front end is achieved by using Millimeter-wave integrated circuits (MIMICs) and planar multi-chip module (MCM) design and process. A traveling-wave feed scheme is adopted to simplify distribution of the local-oscillation (LO) signal. The results show the noise figure of the receiver is less than 5.3dB at 4 GHz bandwidth and the difference of gain of each channel is less than 1dB. The receiver is used in a prototype PMMW imaging system and field tested. The PMMW image of typical scenario is acquired. This works provide a feasible solution schematic for integrated receiver in W-band PMMW imaging system.

Traditional mechanically tunable imaging spectrometry, including those based on MEMS technology, can only obtain continuous spectral image data in a fixed spectral order, which was restricted by the moving mechanism of the mechanical device. But spectrum absorption peakof materials is not continuous, especially in hyperspectral scale. To overcome such shortages, a smart spectral imaging detection method based on the integration of electrically tunable liquid-crystal (LC) Fabry-Perot (FP) microstructure array and a non-continuous wavelength choosing strategy to operation the device are proposed. With a tunable LC-FP array composed by many working units arranging in two dimensions, the device can image in hundreds of spectral bands simultaneity within milliseconds in a calibrated system, and making correct detection for certain objects in one shot.

A combined spectral curve is measured by the traditional field spectrometers in a narrow or wide field range, and the traditional spectral curves are mixed inherently because of the low spatial resolving power. So the mixed spectral pixels were used in the traditional works on the mixed pixel decomposition. A new measure method is proposed to replace the point measurement. The spectral data is captured by the dispersal unit, and one spatial data is from the imaging CCD, and another is captured by the scanning mirror. By the new imaging spectrometer, the target image and spectral curves of every image pixel are captured simultaneously with high spatial resolution. The captured spectral curves could be regarded as the pure curves, which are very useful for the model foundation and analysis of mixed pixel and pixel decomposition. The basic dispersal principle and imaging mode is introduced in this paper. The detailed design is given by the general diagram and the composition figure. The methods and flow chart of the geometric correction and radiometric calibration are discussed in detail. Some experiments are carried out to measure the vegetation and other typical targets, the results of the images and spectral curves are given and discussed.

The problem that the infrared radiation transmits in the atmosphere is always paid attention by the designers who deal with the photoelectricity measurement system and the researchers who deal with the radiation transmission in the earth atmosphere. There is more and more attached importance to the research of the atmospheric transmission model. There are many studies relating to the atmospheric transmission characteristic of infrared radiation, such as the object infrared detection, the navigation, the communication, the earth resource remote survey, etc. The radiation transmittance will be weakened because of the absorption and the scattering. The atmospheric transmittance is an important parameter to measure the attenuation. This paper summarizes the infrared radiation's transmission process in atmosphere, introduces the keystone of atmospheric transmission computation, analyzes the absorption attenuation of the atmospheric gas such as steam and carbon dioxide, the scattering attenuation and the attenuation caused by meteorological condition, and gives the atmospheric transmittance simplified algorithm combining the fact of project. Finally this paper gives the atmospheric transmittance computation result under given meteorological condition, also gives the contrast with LOWTRAN7.0. The algorithm reduces the troublesome step brought by using atmospheric transmission software, it can basically satisfy the request from engineering calculation without high accuracy although the precision is relatively bad.

In order to improve signal-to-noise ratio (SNR) and image quality, this paper introduces a wavelet-based multiscale anisotropic diffusion algorithm to remove speckle noise and enhance edges. In our algorithm, we use the tool of wavelet to construct a linear scale-space for the speckle image. Due to the smoothing functionality of the scaling function, the wavelet-based multiscale representation of the speckle image is much more stationary than the raw speckle image. Noise is mostly located in the finest scale and tends to decrease as the scale increases. Furthermore, a robust speckle reduction anisotropic diffusion (SRAD) is to be proposed and we perform the improved SRAD on the stationary scale-space rather than on the rough speckle image domain. Qualitative experiments based on a speckle Synthetic aperture radar (SAR) image show the elegant characteristics of edge-preserving filtering versus the traditional adaptive filters. Quantitative analyses, based on the first order statistics and Equivalent Number of Looks, confirm the validity and effectiveness of the proposed algorithm.

The modeling and simulation of target laser scattering characteristic is one of the most important parts of the research of space object optical characteristics, and also is the basement of the design of active detecting system. LRCS is the integrated description of target laser scattering characteristic. In this paper, the BRDF model of space object surface materials, which is compatible with the active detecting geometry, is analyzed and selected, then a visualization method of calculating the LRCS based on OpenGL is put forward. Two 3D space object models composed of different materials are built and the LRCS values at different attitude angles and solar panel rotating angles are calculated. This method can increase the reliability and fidelity of the estimation of target LRCS. The maximum rang of active detecting system is also analyzed based on the estimation of LRCS. These research results of the LRCS can provide reference and foundation for the design of active detecting system.

Imaging guiding and machine vision in the infrared (IR) of military targets on land backgrounds are an extensively studied subject. The railway scene as an important traffic infrastructure usually plays a decisive role in land wars. Real IR images are expensive to obtain, and many researchers cannot afford them. The solution to this problem is generating realistic images at various conditions in computer. In this paper, a physics based model of infrared image synthesis for railway scene is proposed. A method for generating a thermal image of railway scene obtained by a forward-looking infrared (FLIR) sensor is described here. It consists of an integrated process based on thermal model of railway, atmospheric transmission and IR sensor effect. At first, a simple numeric model is proposed to calculate the temperature distribution of railway scene based on theoretical analysis of heat transfer. The typical structure of a railway can be divided into two main components. To simplify the thermal model, two parts of railway are processed independent. The distribution of track section temperature is considered in detail and not considered conduction along the track line. The ballast base is discretized into one-dimension multiple layers. Then we focus on the emissivtiy of steel track, which is a dominant factor in railway IR simulation. The value of emissivtiy is mainly determined by surface status of track. So the infrared radiation from track surface is calculated by Stephan-Boltzmann Law. Some results of synthesis image of railway scene in atmospheric window are shown finally. The generating images of railway are in good accordance with real images.

Satellite multispectral imageries are usually limited in low space resolution, long revisit cycle or high cost. This paper presents our ongoing research on developing cost-effective unmanned airship on board Multispectral imagery system to acquire high-resolution multispectral imagery for quick-response to drinking water pollution issues. First, the overall architecture of developed system is described. After that, system integration including CCD cameras coupling, GPS/INS synchronization, stabilize platform control and wireless communication are discussed in detail. Next, system calibration is implemented in radiance and geometry respectively. An adaptive calibration method is developed to obtain absolute radiance and classic homography principle is employed to relate CCD cameras with each other geometrically. Finally, flight experiments are implemented to acquire high-resolution multispectral imageries along river and imageries are deliberately calibrated for the estimation of water quality. Conclusions are also conducted as well.

A dual-waveband infrared re-imaging system has been designed with less lens number and short system length. The properties of re-imaging system to correct chromatic aberration have been researched. The main parameters of this system include F number, focal length, diameter, dual waveband, and imaging plane size are 2,000 mm, 200 mm, 3-5 μm /8-12 μm, and 9.3 mm × 12.4 mm, respectively. After design the average modulation value of all fields of view is 0.60 at 10lp/mm, and the difference between limited diffractive system and actual system is less than 0.3. The ray aberration has been corrected at the center wavelength 5 μm. The focal length varies with wavelength around 400 mm, and gets 400 mm exactly at 4 μm and 10 μm.

The convex grating is one of the key elements in Offner hyperspectral imaging spectrometers. In this paper the design of the holographic convex grating focuses on the optimization of its recording parameters for aberration correction. Meantime, the diffraction efficiency of the convex grating is discussed by using rigorous coupled-wave theory. The results show: within the wavelength range from 0.4μm to 0.78μm, the diffraction efficiency can be over 20% and up to 40% through controlling the groove depth and duty cycle of rectangular grating; the diffraction efficiency can be over 35% through controlling the blaze degree of the blazed grating. The rectangular surface-relief convex grating with the period of 5μm in the center, and the ruled area - a convex with its radius 100mm and aperture 40mm has been fabricated by holographic-ion beam etching. A prototype blazed convex grating will be fabricated in late 2010.

A new multi-spectral image super-resolution reconstruction algorithm based on B-spline interpolation is presented. In order to take the advantages of both multi-frame and single-frame SR methods, B-spline interpolation SR method is proposed by considering sub-pixels of the images which have half-pixel difference with each other. Four B-spline curves are formed by the 12 points which are in the scope of the interpolation point, it take the mean of the midpoints as the value of the interpolation point. Furthermore, uniform B-spline has convexity-preserving property, the curve doesn't interpolate any control points, while two new control points are inserted between two adjacent control points, the points merge the old points to form a new set of control points. The B-spline curve generated by the new set of points includes all old control points, and the interpolating curve is convexity-preserving. This B-spline interpolation SR method can remove blurring and noise produced by the finite detector size and optic and add the image detail. The effectiveness of the proposed method is verified by comparing it against some conventional methods.

Decision fusion can be defined as the process of fusing information from individual data sources after each data source has undergone a preliminary classification. In this paper, a combination of multi-level neural networks decision fusion schemes will be tested in classification of multisource and hyperdimensional data sets. The integrated features of the multispectral image to classify image's texture is used, namely, the two types parameters are estimated as the texture features: the Hurst parameter and the unit displacement incremental power. The efficiency of the features is evaluated by comparing several other features with them. The performance of the above approaches with the use of different feature was investigated. The algorithm presented in the paper was found to be more efficient than other spatial methods.

This paper introduces a novel idea, innovative technology in building multi spectral imaging based device. The benefit from them is people can have low cost, handheld and standing alone device which makes acquire multi spectral images real time with just a snapshot. The paper for the first time publishes some images got from such prototyped miniaturized multi spectral imager.

Orientated towards the application of ship detection in SAR imagery, several typical distributions used for describing the single look synthetic aperture radar (SAR) amplitude imagery and their parameter estimation methods have been summarized. Using the histogram fitting method, this paper analyzes the SAR sea clutter's statistical characteristics and carries out some statistical modeling experiments based on numerous measured SAR images. The results indicate that the log-normal and amplitude-K distributions are suitable to model the experimental data among the several distributions. These two distributions have almost the same modeling precision, but the former one's calculation is more efficient.

Satellite orbital state vectors are required not only to determine the baseline parameters, the coarse offset in registration step but also to remove the phase of the flat Earth and geocode the InSAR products to WGS84. The quality of orbit plays an important role in InSAR products, i.e. DEMs and deformation maps. The orbit inaccuracy in along-track, radial and across-track directions is transformed into a noise baseline vector. Baseline errors are introduced into the InSAR products in the Flat-Earth phase subtraction and topographic phase subtraction processing steps. Under the transformation, the influences of the orbit inaccuracy in the two InSAR processing steps and the InSAR products are analyzed. According to the error propagation theory, we built the relationship between orbit errors and InSAR products. The influence of orbit errors caused by imprecise topography subtraction is small in comparison to imprecisely subtracted flat-Earth phase. The propagation of orbit errors to deformation maps is derived, revealing that the influence of orbital error on deformation is much lower than that on height and only sub-millimeters.

Image fusion is an important technique to integrate high resolution panchromatic image and low resolution multispectral image. The fused image enhances the capability for image interpretation. EMD (Empirical Mode Decomposition) is an effective method to decompose the linear or nonlinear signals into a set of IMFs (Intrinsic Mode Functions). The characteristics of EMD may apply to image fusion technique. The objective of this investigation is to establish a novel image fusion method using a 2-D EMD. The idea of image fusion based on EMD is to decompose the panchromatic and multispectral images into their IMFs. Then, we replace the high frequency IMF of multispectral image by high frequency IMF of panchromatic image. Finally, the image fusion is performed by reconstructing the mixed IMFs. The experimental results indicate that the proposed method may produce a fused image that preserves spatial and spectral information.

The coefficient of variation (CV )[1] and the ratio of arithmetic to geometric mean( A/G )[2] are two methods of heterogeneity measurement for SAR image, they are widely used in SAR image processing. Lopes pointed out that the behavior of the traditional adaptive filter (Lee[3], Frost[4], etc) using for speckle was under the control of the coefficient of variation in essence, and Lopes also deduced the enhanced adaptive filter algorithm. The filter algorithm for speckle based on the A/G is immature. In this paper, A/G is used to replace the coefficient of variation and applied into the enhanced adaptive filters. The novel class of adaptive filters, which are based on A/G and are designed for speckle, are proposed. Through simulation, we verify the effectiveness and superiority of the novel class filters.

The difficulties of BJ-1 image fusion are the big resolution gap (up to 1:8) between the multispectral and panchromatic image and the task to seek a balance between high spatial resolution and the least spectral distortion. In this paper, an experiment has been accomplished prior to the determination of fusion models with a conclusion of the close relation between approximation and spectral distortion. Hence, a combination method of IHS and wavelet decomposition is proposed, of which an optimal fusion model based on spectral and spatial statistical indexes is designed for approximation coefficients in an effort to compromise between high spectral preservation (low distortion) and spatial definition. As to the detail coefficients, a set of multi-scale diverse local algorithm inspired by some successful ARSIS models is employed, which includes the adjustment and establishment of relationship between components from multispectral and panchromatic wavelet decomposition. The fusion results are subsequently compared with counterparts of other methods such as IHS, wavelet and PCA. Several evaluating indicators are employed to give quantitative assessment to the fusion performance. Among all the fusion methods, the statistics-based IHS-Wavelet method fusion demonstrates the most satisfactory spatial definition while keeping the least spectral distortion.

Synthetic aperture radar (SAR) provides a powerful tool for forestry inventory because of its all-weather and all-day capabilities. In this paper forest mapping method using bi-aspect polarimetric SAR data acquired from ascending and descending path has been studied. Zhazuo forest farm in Guizhou province was selected as test site and an 8-temporal field experiment was designed to obtain bio-physical parameters and spatial structure parameters of the 12 sample plots. Then the Michigan Microwave Canopy Scattering model (MIMICS) was employed to analyze the seasonal variation of these 4 types of managed forests. Using polarimetric Radarsat 2 data, scattering mechanisms of each forest type were determined and polarimetric variables were extracted and analyzed for forest discrimination. Considering the inherent geometric distortion of SAR imaging in hilly areas, a geometric correction strategy using bi-aspect SAR images and high resolution DEM was proposed. Then support vector machines method was adopted for classification of the whole test area. Experiments show that the bi-aspect geometric strategy is useful for hilly areas especially for shadow elimination in SAR image, and polarimetric SAR data is helpful to forest mapping.

To get high-precision attitude of the spacecrafts in high dynamic maneuver, a double-FOV(field of view) star sensor is proposed. It is composed of two lenses and two image sensors. The star images of the smaller FOV are used to estimate the attitude of the spacecraft for high-precision, and those of the larger FOV are used to speed the star identification process. Moreover, when the smaller FOV can not capture enough stars, the stars in the larger FOV can be used to estimate the attitude of the spacecraft, which enhances the robustness of the star sensor. The simulation results show that the star sensor can provide high-accuracy attitudes while the spacecraft is in high dynamic maneuver stage.

Normalized difference vegetation index (NDVI) is defined as a ratio of the difference of the infrared and red bands to the sum of the two bands. It can be estimated directly from satellite data, and has been widely used in numerous environmental studies. Yet the satellite-based NDVI was criticized for its variations with temporal factors (e.g. sun-surface-satellite geometry, atmospheric variations). Such variations may result in false change of vegetation over surface. However, the uncertainties relevant to the false change are generally unquantified in the studies. It is therefore unclear to what extent the satellite-based NDVI would be reliable. In this study, we used a derived relationship between the digital number (DN) with and without temporal influences for the same area. Using the derived relationship, NDVI can be expressed as a function of atmospheric optical thickness (AOT), view angle, and DN without temporal influences. As a result, the uncertainties relevant to the temporal factors were quantified with a mathematical expression. We found that satellite-based NDVI was a function of AOT, day of year, latitude, and NDVI without temporal influences. We made simulations in the case of Landsat TM data. Simulations showed that atmospheric effect was most influential to a satellite-based NDVI, and the NDVI would suffer more serious influences at higher latitude than at lower latitude. In general, the temporal influences on NDVI cannot be ignored for a reliable monitoring of surface phenological processes.

In this paper, we propose a method based on both 3D-SPECK (3D Set Partitioning Embedded Block) and the theory of DSC (Distributed Source Coding) to realize the compression of hyperspectral images. Some experiments have been done to evaluate the compression performance, and the experimental results show that our methods can get competitive performance than 3D-SPECK, 3D-SPIHT.

Based on the infrared radiation characters of outer space targets and environment, considering the technologies related to the infrared star grade, atmosphere attenuation, and coordinate transformation, this paper builds the theoretical model of infrared radiation from flying satellites and background. Furthermore, using the database management of satellites orbits and catalogue data, the dynamic scene image synthesis of satellites and background is implemented. Finally, we can walkthrough the virtual scenes from different viewports, and analyze the characters of these dynamic simulating images. Research results have significance for space targets exploration, identify and tracking.

With the availability of high resolution multispectral imagery from sensors, it is possible to identify small-scale features in urban environment. Given attributes of image structure such as color, texture, have the character of highly scale dependency, a hierarchy segment fusion algorithm based on region deviation is proposed to extract more robust features and benefit single semantic level land cover classification. The fusion algorithm proposed is divided into in two successive sub-tasks: mean shift (MS) filtering based pre-segmentation and hierarchical segment optimization. Presegmentation is applied to get boundary- preserved and spectrally homogeneous initial regions, and then, a family of nested image partitions with ascending region areas is constructed by iteratively merging procedure. In every scale, regions of the corresponding critical size are evaluated according to potential region merge risk, which is measured by the region standard deviation change before and after a virtual merge. If a region measurement is larger than a specified change threshold, the region will be preserved to the next level and labeled as a candidate segment for following regionbased classification. Otherwise the segment will be merged to the next scale level. After fusing segments in different scales, a novel weighted minimum distance classifier is employed to get supervised classification result, in which every feature band's deviation is used to calculate its own weight. We show results for classification of a HR image over Washington DC Mall area taken by the HYDICE sensor. Different features combined with designed classifier have proved that fused segments provided a robust feature extraction and improve classification accuracy.

Image registration is a vital step in the processing of multispectral remotesensing imagery. This paper presents a robust multispectral remotesensing image registration algorithm based on maximally stable extremal regions (MSERs). Firstly, MSERs are detected independently in the reference image and the sensed image. Secondly, the SIFT descriptor is adopted to capture texture information in the detected regions, while an affine invariant shape descriptor for MSER is constructed to ensure that features can be reliably matched regardless of the appearance change. Both the SIFT descriptors and the shape descriptors are matched using the Euclidean distance measurement. Matching results are then combined and the optimal corresponding points are chosen to estimate the transformation parameters. Finally, random sample consensus (RANSAC) algorithm is applied for geometry estimation. Experimental results on various image pairs demonstrate that the proposed MSER based algorithm is very effective for multispectral remotesensing image registration.

Cloud segmentation is an important step and a very difficult problem in typhoon image processing. There are many works on cloud image segmentation, but few are carried out on typhoon primary cloud system (galaxy) segmentation. Typhoon satellite images are always multiple channels whose properties are very different, so that the appearances of these channels are different as well. In order to segment out primary cloud systems accurately, multiple channel images are employed in this paper. The image data is from MERSI (short for MEdium Resolution Spectral Imager) of Chinese FY- 3A meteorological satellite launched on March, 2008. The scalar multiphase Chan-Vese (CV) model is extended for the vector-valued images, so as to partition out typhoon cloud systems. The experiment results show that the multi-channel segmentation is more accurate, more complete and more effective than that of usually using only one image, with multiple channel images being treated as a vector one input into the CV model. The multi-channel segmentation integrates the distribution information of cloud systems in all channels, so information fusion of multiple channels are realized when segmenting.

An improved Nonsubsampled Contourlet Transform (NSCT)-based method has been proposed in this paper, using subbands mask prior models and directional information for synthetic aperture radar (SAR) image despeckling. The NSCT is a shift-invariant directional wavelet transform which satisfies the demands of SAR image despeckling. Wedding coefficient statistic character of NSCT subbands and directional information captured by the transform, this paper applies anisotropic neighborhood models to the processing of adaptive shrinkage. The resulting algorithm strongly suppresses speckle, while preserving detailed information and image edges compared with other traditional despeckling methods.

In this paper, we proposed a SAR speckle reduction method based on sparse and redundant representations over multiscale ridgelet dictionary. Firstly, the multiscale ridgelet function is proposed. And then based on it, the multiscale ridgelet dictionary is constructed, which can sparsely represent the SAR images. Finally, we propose a global image prior that forces sparsity over small patches in every location in the image. We define a maximum a-posteriori probability (MAP) estimator as the minimizer of a well-defined global penalty term. The speckle reduction leads to a simple iterated patch-by-patch sparse coding and averaging algorithm. The experimental results demonstrate that the proposed method performs better than several other existing methods in terms of quantitative performance as well as in term of visual quality of the images.

Multispectral images are available for different purposes due to developments in spectral imaging systems. The sizes of multispectral images are enormous. Thus transmission and storage of these volumes of data require huge time and memory resources. That is why compression algorithms must be developed. A salient property of multispectral images is that strong spectral correlation exists throughout almost all bands. This fact is successfully used to predict each band based on the previous bands. We propose to use spectral linear prediction and entropy coding with context modeling for encoding multispectral images. Linear prediction predicts the value for the next sample and computes the difference between predicted value and the original value. This difference is usually small, so it can be encoded with less its than the original value. The technique implies prediction of each image band by involving number of bands along the image spectra. Each pixel is predicted using information provided by pixels in the previous bands in the same spatial position. As done in the JPEG-LS, the proposed coder also represents the mapped residuals by using an adaptive Golomb-Rice code with context modeling. This residual coding is context adaptive, where the context used for the current sample is identified by a context quantization function of the three gradients. Then, context-dependent Golomb-Rice code and bias parameters are estimated sample by sample. The proposed scheme was compared with three algorithms applied to the lossless compression of multispectral images, namely JPEG-LS, Rice coding, and JPEG2000. Simulation tests performed on AVIRIS images have demonstrated that the proposed compression scheme is suitable for multispectral images.

Space-time Information Expression and Analysis (SIEA) uses vivid graphic images of thinking to deal with information units according to series distribution rules with a variety of arranging, which combined with the use of information technology, powerful data-processing capabilities to carry out analysis and integration of information units. In this paper, a new SIEA approach was proposed and its model was constructed. And basic units, methodologies and steps of SIEA were discussed. Taking China's coastland as an example, the new SIEA approach were applied for the parameters of air humidity, rainfall and surface temperature from the year 1981 to 2000. The case study shows that the parameters change within month alternation, but little change within year alternation. From the view of spatial distribution, it was significantly different for the parameters in north and south of China's coastland. The new SIEA approach proposed in this paper not only has the intuitive, image characteristics, but also can solved the problem that it is difficult to express the biophysical parameters of space-time distribution using traditional charts and tables. It can reveal the complexity of the phenomenon behind the movement of things and laws of nature. And it can quantitatively analyze the phenomenon and nature law of the parameters, which inherited the advantages of graphics of traditional ways of thinking. SIEA provides a new space-time analysis and expression approach, using comprehensive 3S technologies, for the research of Earth System Science.

Two-spectrum (LW/MW) IR imaging system using Automatic Target recognition technology, which has both merits of single long and middle IR wave, can performs well in clutter sea-sky condition. The information processing technology of dual-band IR imaging guidance plays a key role in the research of dual band IR system using ATR technology. In the paper, we present a system for ATR technology on dual-band IR imagery. We described the system architecture and methods for target detection, recognition and tracking.

Land surface temperature (LST) is an important measurement for estimating equilibrium of income and expense of land surface energy. It is also a key input parameter in many geographic models. Therefore, research on land surface temperature retrieval has close relation with thermal infrared-related study, such as hydrology, ecology, climatology, environment and other fields. Made in China, the Small Satellite Constellation for Environment and Disaster Monitoring and Forecasting is an advanced satellite constellation (composed of satellite HJ-1A, 1B and 1C) designed for environment and disaster monitoring and mitigation. Whether the sensor data can reach the designed specifications and meet the demands of application? It is necessary to carry out relative research before the launch of a new satellite. There is an infrared sensor in HJ-1B. Our work has been done before the launch of HJ-1B. This paper focuses on the land surface temperature retrieval study based on HJ-1B thermal infrared data, which is significant for its potential assessment and effective application in environment monitoring and disaster preventing and management. According to the characteristics of HJ-1B thermal infrared sensor, a method of using middle infrared (MIR) band and thermal infrared (TIR) band of HJ-1B is put forward in this paper. The spectral response function of bands, standard atmospheric profiles data and radiation transfer simulating software-MODTRAN are used to get simulated HJ-1B infrared data. And finally, the algorithm accuracy is estimated by comparing the retrieval value and true value of temperature. And the sensitive analyzing of retrieval algorithm is made through some main parameters. It can be know from our research that the proposed land surface temperature retrieving algorithm for HJ-1B infrared data has a considerable precision, the RMSE value range is 0.01K~2.08K. The RMSE increases with the increase of view zenith angle. The variation range of temperature retrieval RMSE due to view zenith angle is 0.1K~0.2K. The emissivity and water vapor content influence the land surface temperature retrieving result obviously, and the influence of instrument noise on retrieving result is little and can be ignored.

Registration of two or more complex images of the same scene is an important procedure in InSAR (Interferometric Synthetic Aperture Radar) image processing. Meanwhile, the accuracy of this step is crucial to the reliability of subsequent image processing and final results of the data processing chain. This paper presents a robust method which introduces the coarse-to-fine strategy, relaxation optimization technology and Maximum Spectrum Method to obtain the dense, reliable and accurate conjugate points for the registration of two single looking complex (SLC) SAR images. In each pyramid image, we extract the feature points using the Moravec interest operator, and then perform matching for feature points to find their candidate conjugate points whose correlation coefficient is above certain threshold, and through relaxation technology to find the best matcher. The matching result at higher pyramid level is then used to guide and limit the search space for the matching in the lower level. Perform above procedure iteratively until to the original image, the Maximum Spectrum Method is carried out to refine the matching accuracy to the sub-pixel. After determining dozens of thousands conjugate points on master and slave image, we can form the transformation model between them and perform image correction. We have made experiments to validate our method, and it comes to the conclusion that with coarse-to-fine strategy, global relaxation and Maximum Spectrum Method, it efficiently reduces the matching error and improves matching precision.

Local invariant features such as scale invariant feature transform (SIFT) have received considerable attention in recent years. Despite its tremendous success in computer vision applications, SIFT matching alone is not sufficient for remote sensing image registration because of low detection repeatablility and nonlinear intensity changes. In this paper, we introduce a remote sensing image registration algorithm that combines local affine frames (LAF) together with SIFT matching. Firstly, distinctive SIFT keypoints and maximally stable extremal regions (MSER) are detected independently in the reference image and the sensed image. Contrast reversal invariant SIFT descriptor is constructed for describing texture patches around SIFT keypoints and shape descriptor defined in LAF is constructed for describing MSER contour. Nearest neighbor distance ratio matching with confidence measurement is then adopted to match both descriptors. Tentative correspondences are ranked according to their confidence measurements. Finally, random sample consensus (RANSAC) is performed in the top ranked matched features to obtain a global set of transform parameters. Experimental results demonstrate the robustness and accuracy of the proposed method.

In this paper, we propose an Infrared(IR) background simulation method by integrating texture modeling and infrared prediction. First, by introducing the latest work of compute vision, we argue that the infrared texture play more important role of scene configuration contrast to the traditional viewpoint that the infrared texture is just used to overcome low resolution of model or for feature enhancement. Next, we present the infrared texture should be simulated in the radiant energy space relative to the temperature field, and synthesize the infrared texture using FRAME model. In the end, according to the scene model introduced from the compute vision context, we present the IR background simulation method by integrating the IR prediction of material region and the corresponding texture synthesized by FRAME model.

Remote sensing image classification is a prerequisite for remote sensing applications, such as thematic mapping, urban planning, forest management, environment monitoring, disaster warning and assessment, military target recognition. Over the last decade there has been noticeable shift in remote sensing image classification with the extension of remote sensing imagery sources as well as the development of pattern recognition methods. This paper discusses the changes in remote sensing classification from two aspects: basic thought and new classification algorithms. The basic thought of remote sensing classification has changed from per-pixel multispectral-based approaches to multiscale object-based approaches. New classification algorithms include support vector machine, evolutionary algorithm, fuzzy clustering algorithm, as well as Artificial Neural Networks. At last this paper highlights the future research and application directions of remote sensing image classification.

This paper presents the simulations of TWRI (through the wall radar imaging) in 2D and 3D scene and discusses the finite beamwidth processing to reduce the high calculation of the back-projection algorithm. The 2D and 3D scene are reconstructed to display the performance of the TWRI system using analysis, numerical simulations and real data collected from Through-the-Wall Radar (TWR). At the end of this paper, a novel approach is proposed to locate the position of the target accurately using the Ultra-Wideband Short-Pulse (UWB-SP) Radar.

In this paper, a new multi-spectral remote sensing image segmentation method based on multi-parameter semi-supervised spectral clustering (STS³C) is proposed. Two types of instance-level constraints: must-link and cannot-link are incorporated into spectral cluster to construct semi-supervised spectral clustering in which the self-tuning parameter is applied to avoid the selection of the scaling parameter. Further, when STS³C is applied to multi-spectral remote sensing image segmentation, the uniform sampling technique combined with nearest neighbor rule is used to reduce the computation complexity. Segmentation results show that STS³C outperforms the semi-supervised spectral clustering with fixed parameter and the well-known clustering methods including k-means and FCM in multi-spectral remote sensing image segmentation.

To implement multi-frame super-resolution restoration of low-resolution images or video sequences with nonglobal motion, an interpolation-filtering method which is based on the recently proposed nonlocal-mean (NLM) filter is presented. Firstly, each frame of a sequence is interpolated to desired resolution by cubic-spline or other algorithms. Then the NLM filter is applied on these interpolated frames, while this filter is extended from single image at spatial domain to multiple images at spatial and temporal domain simultaneously. Finally, these high-resolution images are used to reconstruct the anticipant images or sequence. Although the NLM does not strongly rely on the accurate motion estimation and is not sensitive to nonglobal motion, to reduce the complexity and improve results, a two-step method for coarse motion estimation is adopted to obtain the motion vectors between frames. The obtained vectors act as initial candidate location for the weight computing and pixel filtering of NLM. The performances are tested on a simulated data and on a real video sequence together with the existing methods. Results on these tests show that the proposed technique is successful in providing super-resolution on general real sequences.

Usually, Wishart H/α/A classification is an effective unsupervised classification method. However, the anisotropy parameter (A) is an unstable factor in the low signal noise ration (SNR) areas; at the same time, many clusters are useless to manually recognize. In order to avoid too many clusters to affect the manual recognition and the convergence of iteration and aiming at the drawback of the Wishart classification, in this paper, an enhancive unsupervised Wishart classification scheme for POLSAR data sets is introduced. The anisotropy parameter A is used to subdivide the target after H/α classification, this parameter has the ability to subdivide the homogeneity area in high SNR condition which can not be classified by using H/α. It is very useful to enhance the adaptability in difficult areas. Yet, the target polarimetric decomposition is affected by SNR before the classification; thus, the local homogeneity area's SNR evaluation is necessary. After using the direction of the edge detection template to examine the direction of POL-SAR images, the results can be processed to estimate SNR. The SNR could turn to a powerful tool to guide H/α/A classification. This scheme is able to correct the mistake judging of using A parameter such as eliminating much insignificant spot on the road and urban aggregation, even having a good performance in the complex forest. To convenience the manual recognition, an agglomerative clustering algorithm basing on the method of deviation-class is used to consolidate some clusters which are similar in 3by3 polarimetric coherency matrix. This classification scheme is applied to full polarimetric L band SAR image of Foulum area, Denmark.

The detection and tracking of dim moving targets in very low signal-to-noise ratio (SNR) environment has been a difficult problem in radar signal processing. For low SNR moving targets detection, a new improved dynamic programming algorithm based on track-before-detection method is presented. This new algorithm integrates energy along target moving tracks according to target moving parameter information. This process substitutes the exhaustive search by a feasible algorithm. The simulation confirms that this algorithm, with high computational efficiency, is feasible, and can effectively estimate trajectories of dim closing moving targets. The process has also been shown to give an increase in detection.

In this paper, we propose a novel change detection method. Multiple classifiers fusion combine results from various simple changes detection methods to improve change detection accuracy. In detail, we make use of multiple classifiers fusion based on fuzzy integrals for change detection. If the fuzzy measures are well defined, the accuracy of change detection can be improved distinctly. In this paper, we determine the fuzzy measures based on the Genetic Algorithm (GA). Though multiple classifiers fusion has robust performance, the input change detection result is still important. We review proposed pre-classification change detection method, and propose two contextual Fuzzy-C Means (FCM) algorithms and the Self Organization Feature Map (SOFM) change detection method. We select multi-spectral TM and pan SPOT image pairs as test data and apply five different change detection methods. The first experiment shows that different methods will produce different change detection accuracy, and different methods will complement each other. In addition, we apply fuzzy integral aided by genetic algorithm for combining different detection methods. The final experiment shows that our proposed method can improve change detection accuracy and has better performance than single detection method.

Gram-Schmidt and Ehlers fusion, which are well known for spectral fidelity, are described. Selecting a sandrock mine using SPOT5 as study area, spectral fidelity and high spatial information gain are used to assess the two fusion methods, which are compared with multiplication and Andorr fusion. In the whole, the Gram-Schmidt method is the best, preserving highly the original spectral information, and can provide spectrum control foundation for interpreting mine targets in the complex geology environment. Ehlers method is the second. Then Andorr method is the third, and it has the highest spatial information gain, but high frequency information is enlarged excessively, effecting on the identification of mine exploitation state. The multiplication method is the worst, because it loses the high and low frequency information, which is the most important for mine targets recognition.

HJ1A HSI is an interferometric imaging spectrometer (Hyper-spectral Imager, sensor ID: HSI) of HJ-1 small satellite. The hyper-spectral image data are organized and stored in hierarchical data format version 5 (HDF5) files. This paper presents the data model, file structure, library and programming model of HDF5 file format. The adapter design pattern is used for translating hdf5 interface into a compatible interface. Then, we give a detailed analysis of HJ1A hyper-spectral image data. The HJ1A hyper-spectral image data model includes five groups: 'GlobalAttributes', 'ImageAttributes', 'ImageData', 'MapInformation', and 'ProductParameters'' under the root group. The 'ImageData' group includes three datasets: 'BandData', 'CalibrationCoefficient', and 'WaveLength'. Based on the relationships between the models and implementations, we give a flow chart of extraction HJ1A hyper-spectral image data from hdf5 files. The level2 product of HJ1A hyper-spectral image data is used for experiment. We present the RGB color composite image and 3D cube of the extracted data. Tests show that the data extraction is correct and rapid with this approach. This work provides a solid foundation for quality evaluation and application of HJ1A hyper-spectral image.

There are some disadvantages existing in traditional artificial method of detecting road rut feature, such as consuming time, consuming energy, low accuracy and danger. In order to solve those problems listed above, a new method of detecting features automatically based on the laser image processing technology is proposed. The first step is to start pretreatments such as fuzzy compensating and enhancing the original images. Then adopting an improved watershed serial segmentation method, we can get binary segmentation images. Used multi-level filtering technology combined with statistical methods, the accurate road rut feature can be got. It has been proved by experiments that the detection method has high accuracy and stability.

In this paper a new dim target fusion detection method based on SA4 multiwavelet is proposed to solve the problem that scalar wavelet has a low precision in dim target detection. Multiwavelets have more freedoms in their construction and thus can combine more useful properties than the scalar wavelets. Firstly sixteen sub-images of different resolution can received by decomposing the image with one-layer multiwavelet, then the four low frequency coefficients of multiwavelet are removed by setting them to zero, and the sub-images are fused with weighted average algorithm, while the optimization assignment of weight can be got using Lagrange Multiplier method.The detection result is received by segmenting with adaptive threshold to the fusion image, and the moving track of target can be got by accumulating the results of each frame. Two sequence images with SNR of 2.71 and 1.81 are used in the experiment respectively, and experimental results show that the contrast of the image processed with multiwavelet is more clear than the scalar wavelet one, and the detection precision is distinctly improved.

To improve the fusion effect of the PAN (Panchromatic) image and the MS (Multi-Spectral) image, this paper proposes a new fusion approach with high fidelity of color based on the combination of IHS and wavelet transformation. After the PAN image and the intensity component of MS image are decomposed respectively by wavelet transformation, lowfrequency sub-images are fused with a self-adaptive fusion rule based on the improved PCA (Principal Component Analysis), while the fusion rule of maximum of absolute values with consistent verification is used to fuse highfrequency sub-images. Then the intensity component of MS image is replaced by the merged data and the fused image is obtained by inverse IHS transform of the MS image. The subjective visual effect and the objective estimation indicate that the new method can preserve the color information very well. Compared to the traditional IHS and wavelet transform based algorithms, the color distortion of the proposed algorithm reduces 47.31% and 26.14% separately. In addition, compared to the weighted average fusion method of low-frequency coefficients and high-frequency coefficients substituted directly by high-frequency sub-images of the original PAN image, the color distortion of the introduced intelligent and self-adaptive fusion rule reduces 6.51% and 24.22% respectively.

Research on the location of sea-sky line under infrared complicated sea-sky background has the important value to improve the capability of far target recognition. In a real sea-sky background infrared image, the sky and the sea normally appear in the form of different changes of grey-level on account of their different infrared radiation and reflection characteristics. So we can use the background complex degree to differentiate between the sky region and the sea region. In our method, an operator is employed to calculate the background complex degree of image. By filtering original infrared image with the operator, the grey-level contrast between the sky region and the sea region can be stronger than original image. After the complex degree calculation, the edge detection with only row directional gradient and Hough transform are utilized in succession, the sea-sky line can be located. Some typical sea-sky background infrared images are selected to validate our proposed method. The experimental results indicate that our sea-sky line location method is robust.

Bandelet transform is an efficient image sparse representation approach which can adaptively approximate the geometrical regularity of image structures. In this paper, a multi-bandelets based method for SAR image compression is presented, which is constructed by combining multi-wavelet with Bandelet transform and geometric flow optimization. Compared with single wavelet, multi-wavelet has some advantages such as compact support, orthogonality, symmetry and smoothness, thus making finite length filtering, linear phase, correlation remove and good frequency domain characteristics possible, which are very desirable in image compression. Moreover, in our method the multi sub-bands collaborative decision algorithm for geometric flow optimization is proposed to obtain more accurate geometric flows. A number of simulations are taken on SAR images and the result shows that our method can provide a significant improvement over the multi-wavelet and the second generation Bandelet, both in visual fidelity and some objective evaluation criteria such as peak signal to noise ratio, equivalent numbers of looks and edge preservation index.

The fusion of panchromatic image and multispectral image is one of the significant issues in the application of remote sensing , by which it can integrate multispectral and panchromatic images with their respective strengths to compress the data size and improve data utilization. In this paper, an IHS (Intensity-Hue-Saturation) image fusion method is presented based on Bidimensional Empirical Mode Decomposition (BEMD).Firstly, multispectral image is transformed form the RGB color space to HIS color space, then I component and panchromatic image are carried out decomposition respectively using BEMD method. The intrinsic mode surface of decomposed panchromatic image is used to replace that of decomposed I component. Finally the image is transformed back to original RGB color space from IHS color space, achieving the image fusion. The experiment shows that the new image fusion method is superior to the traditional methods.

This paper discusses how to separate non-ground points from raw LIDAR point cloud. For the purpose of improving processing efficiency and precision, an improved 1-D filtering method is proposed. The entire filtering process is divided into eight steps and non-ground points are eliminated progressively. In these processing steps, a key-point detection technique is used to segment points in profile. Based on these profile segments, detailed analysis is utilized to implement segment-oriented filtering innovatively. This method makes use of entire features of segmental points for classification, so it is more accuracy and robust than traditional point-by-point classification. Two different scale datasets are used to test our method. Compared to 1-D labeling method, the proposed method is more effective and efficiency.

In this paper, combined with Canny's criteria, we proposed an improved Rothwell edge detection method which aims to recover more reliable topological relations from the extracted edges. Distance transform is used to optimize location of the edge and B-Spline interpolation is applied for the detection of zero-crossing in the sub-pixel interpolation step, and edge refinement is used to adjust topology. Experiments results show our improved Rothwell edge detection algorithm ^[1] could retain more accurate geometry features while generate more stable topology than the original one.

This paper proposes a novel multiscale graph cut based analysis framework for the supervised classification of hyperspectral imagery. This framework is aimed at obtaining accurate and reliable maps by properly considering the spatial-context information. It is made up of two main blocks: 1) a feature-extraction block exploits an object-oriented analysis and representation of hyperspectral imagery that is obtained by multiscale graph cut (MGC) based segmentation; 2) a classifier, based on support vector machines (SVMs), capable of analyzing hyperdimensional feature spaces. Experimental results confirm the effectiveness of the proposed system for the analysis of hyperspectral imagery.

Synthetic aperture radar (SAR) images compression is very important in reducing the burden of data storage and transmission. Finding efficient geometric representations of images is a central issue in improving the efficiency of image compression. Bandelet provides an efficient way for image representation based on geometric regularity. In the second generation Bandelet, the multiscale decomposition of image is completed by 2D wavelet transform (WT) and the obtained subbands images are squared partitioned. Then a bottom to top CART algorithm is used to prune the quadtree, and finally an exhaustive searching algorithm is used to obtain the optimal direction in each square. This process is of high complexity in time and space though it can provide an efficient representation of images than WT. Considering this, we proposed a rapid implementation of Bandelet transform based on fixed size image partition, and then applied it to SAR image compression. Experiments results show that in relative to the second generation Bandelets, our proposed method has rapid implementation and comparable performance with chinalake and abq_apt in 0.5-2.0bpp. An improvement of PSNR(Peak Signal to Noise Ratio) and the preservation of edges and texture over JPEG2000 are obtained.

Because of noise and clutter, the infrared target detection even becomes more difficult. In this paper, we present an automatic seed selection method based on an improved mountain cluster algorithm to be employed in infrared target segmentation. The original image was projected to x and y coordination firstly. Then the modified mountain clustering algorithm is employed on the projection data to obtain the cluster center which is applied as the seed point for region growing. This approach has been performed on a field infrared image. The experimental results show that the proposed method can provide satisfactory segmentation result.

Image fusion is a process of combing multiple images of the same scene into a single image with the aim to preserve the full content information and retain the important features from each of the combined images. In this paper, a novel image fusion method based on Wavelet Transform (WT) and Visual Attention Mechanism (VAM) is proposed. Firstly, the source images are decomposed by WT to get the sub-images. Secondly, by using the VAM, the salience maps of the source images are formed, which can indicate the salient regions to the human visual system, i.e., the higher the saliency value is, the more important the location represent. The saliency maps are then used together with the match measure between the coefficients to guide the combination of the coefficients as follows: if the match measure at a given location is low, the coefficient from the source image with higher saliency is selected to be the fused coefficient. Contrarily, if the match measure at a given location is high, then the fused coefficients are calculated as the weighted sum of the coefficients extracted from both source images. The weights here are determined by the corresponding saliency maps of the source images. Finally, the fusion image is obtained by using the inverse WT transform. Experimental results applying the proposed algorithm show that the fused image keeps more visual meaningful information than other methods.

To further improve the accuracy of crop detection and acquire more information for land use investigation and agriculture management, this paper proposes a variational level set model for crop detection by combining airborne LiDAR(Light Detection and Range) points cloud and aerial image simultaneously acquired by LiDAR device. Specifically, normalized digital surface model (nDSM) derived from raw LiDAR points cloud are combined with aerial image so as to alleviate the misclassification caused by insufficient information only based on remote sensing image data. This fusion combines spectral and height information of objects from both sensors. By classifying the combined image using our proposed level set model, crop can be discriminated. Then, the paper suggests a novel method based on classification to predict crop density in a given scene. Experiments have verified that the proposed scheme really improve the accuracy of crop detection and the effectiveness of the proposed scheme of crop density estimation.

This paper introduces the application of a new technique named Interferometric Point Target Analysis (IPTA) to monitor the land subsidence of urban areas in Suzhou city. It focuses on some points with high coherence to process. It uses a large number of SAR data set to analysis the spatial and temporal characteristics of some interferometric point targets. Firstly, it introduces the phase model of IPTA. Then it emphasizes on three aspects: the problem of selecting the master image; the problem of choosing the interferometric point target candidates; the data processing methods of IPTA technique. Then it can accurately get the history of land subsidence and atmospheric delay phase information.

The paper proposes a combination of DCT and the Dual-Tree Discrete Wavelet Transform (DDWT) to solve the problems in multi-spectral image data storage and transmission. The proposed method not only removes spectral redundancy by1D DCT, but also removes spatial redundancy by 2D Dual-Tree Discrete Wavelet Transform. Therefore, it achieves low distortion under the conditions of high compression and high-quality reconstruction of the multi-spectral image. Tested by DCT, Haar and DDWT, the results show that the proposed method eliminates the blocking effect of wavelet and has strong visual sense and smooth image, which means the superiors with DDWT has more prominent quality of reconstruction and less noise.

This work presents the application of PS-InSAR for monitoring large-coverage and dynamical ground subsidence. For the standard PS-InSAR technique, a large stack of InSAR images is necessary (usually more than 20 images). We present a method for the extraction of precise subsidence velocity from small InSAR image stacks. Using ENVISAT/ASAR data within one year duration differential interferograms with multi-master images are generated. A contour map derived from subsidence product shows a good consistency with leveling measurements. It indicates that using a small stack of SAR images acquired within a short-time is possible for monitor subsidence with a reasonable accuracy and reliability.

Scene-based nonuniformity correction algorithms are widely concerned since they only need the readout infrared data captured by the imaging system during its normal operation. A system based on the neural network algorithm is designed for real-time correction, using the framework of foreground and background. FPGA as the foreground performs the regular nonuniformity correction and blind pixel detection. As the background, DSP monitors changes of the scene and updates the correction parameters according to the analysis of the scene. In order to eliminate ghosting artifacts, an edgedirected learning scheme is used. Via testing, the system is capable of tackling 25 frames per second. The performance of the system is evaluated with real infrared imaging sequences. The results show a more reliable fixed-pattern noise reduction, tracking the parameter drift, and presenting a good adaptability to scene changes.

With an increasing number of images collected every day from different sensors, automated registration of multisensor/ multi-spectral images has become an important issue. Image registration is a very important precondition of image fusion. A wide range of registration techniques exists for different types of applications and data sources, however no algorithm is known that can accurately register multi-source images consistently. A new method of remote sensing images registration is presented in this paper. The development of this new automatic image registration method is based on Fourier transform and mutual information, which use wavelet decomposition technique. The result of experiment shows that this algorithm can effectively solve the image registration of remote sensing images with quite different diversities, and it has a high value in use.

The particular noise in image not only lower the quality of image but also make denoising more difficult. The multiwavelet with the orthogonality and symmetry decompose the image signal more precisely and denoise better than wavelet. But for multiwavelet transform, the choice of prefilter is extremely important, meanwhile leads to the complication of computation. Furthermore, the image decomposition based on balanced orthogonal multiwavelet is excellent, and does not need pre-filtering. So, in this paper, we propose a method of combining threshold-selecting model with semi-soft thresholding function for image denoising based on the coefficients of balanced orthogonal multiwavelet transform. The simulation results show that this method is superior to others.

Parametric edge detector has been reported to be successfully applied in actual coherent ladar intensity images corrupted by speckle. But there are isolated erroneous test points in the image processed by parametric test detector. By morphological filter, most isolated missing pixels in edge image background can be removed. Thus, the error of the edge detection can decrease exceedingly which is useful for subsequent image processing. Edge detection plays an important role in ladar image processing. Its capability has direct influence on precision and performance of ladar imaging system. Therefore it is necessary to evaluate the performance of a detector. In order to demonstrate the advantages of morphological filter based on parametric edge detection, the probability that an image pixel may be marked as an edge by this detector is calculated. This detector is applied to detect a large number of images degraded by speckle with different patterns and different carrier-noise-ratio. From the simulation results, the performance of morphological filter based on parametric edge detection is described.

An adaptive SAR image enhancement method is presented for reducing the speckle noise and increasing the contrast of synthetic aperture radar (SAR) images. First, a fuzzy logic based filter, employing fuzzy edge to weight the contributions of pixel values in filter window, is used to filter the speckles. Second, the original SAR image is decomposed into lowfrequency component and high-frequency component. The fuzzy filtered image is viewed as the low-frequency component, and the contrast limited adaptive histogram equalization algorithm is used to increase its contrast. The highfrequency component is obtained by subtracting the low-frequency component from the original image, and its gain is controlled by fuzzy structural which employed to express the degree of a pixel belonging to structures. After processed one after the other, the two components are added together to form the final enhanced SAR image. Experimental results show the excellent effect of the proposed method by visual observation and numerical measurement. Many fine structures and little speckle noise can be seen from the enhanced SAR images.

Vignetting effect is a kind of typical nonlinear effect in the thermal imaging system and it will induce the central region of the infrared image is bright, while the edge region is dark. This geometric phenomenon generates because the radiation illumination which reaches the detector's surface decreases gradually with the increase of off-axis distance. Vignetting makes infrared imaging system receive uneven effective energy of incident light-ray from different angle, and the output signal is uneven sequentially. At last the final infrared image brings undue light or dark distortion and the generated infrared image is inconsistent with the real scene. Infrared imaging system works under low contrast between 1 percent and 2 percent. Therefore, the vignetting effect of thermal imaging system influences the quality of infrared imaging seriously. So the exact modeling of vignetting effect is vital for generating an infrared simulation image through optical system. This paper builds a realistic model for the vignetting effect resulted from infrared optical system and analyses the cause of the vignetting effect theoretically. Then simulation is carried out and the simulated results show that the simulation method of vignetting effect can provide the more precise and real infrared image signal to evaluate the capability of infrared imaging system and advance the whole performance.

This paper presents a new reversible transform of inter-spectrum adjustable parameter matrix, which has better redundancy elimination effect by adjusting magnitude parameter λ and shift parameter δ to adjust transform matrix. Intraframe redundancy is eliminated by integer discrete wavelet transform (IDWT).These two kinds of transforms are all completed entirely by addition and shift, whose fast operation speed makes hardware implementation easier. After interspectrum and intra-frame transform, the hyper-spectral image is coded by improved EBCOT algorithm. Using hyper-spectral images Canal shot by AVIRIS of American JPL laboratory as test images, the experimental results show that in lossless image compression applications the method proposed in this paper is much better than the research results of MST, NIMST, a research team of Chinese Academy of Sciences, DPCMARJ, WinZip and JPEG-LS. The condition in which λ=7 and δ=3 in this paper, on the average the compression ratio using this algorithm increases by 11%, 15%, 18%, 31%, 36%, 38% and 43% respectively compared to the above algorithms. From the foregoing it follows that the algorithm presented in this paper is a very good hyper-spectral image lossless compression coding algorithm.

Noise estimation is an important task in hyperspectral remote sensing image process. However, hyperspectral image has limitation on spatial resolution. Usually, More than one type of materials are embedded in ground sampling distance(GSD) of the image. Therefore, this paper focuses on the evaluation on the effect of spatial resolution on noise estimation. A series of simulative images with different spatial resolution levels are generated for the evaluation. Noise standard deviation, the normalized eigenvalues of minimum noise fraction (MNF) transformation and the endmember extraction accuracy are used to evaluate the noise estimation methods. In experiments, the results demonstrate that the methods who estimate noise in spectral domain have robust stability of spatial resolution and perform better in low spatial resolution levels.

Building recognition is an important field in computer vision. Building target line features which represent the target geometry information are stable features in infrared images. In this paper, the stable building line features are acquired by morphology filter algorithm and the number of the correctly obtained line features and the distance similarities of these features are used to measure the probability of the target emerge in the infrared image. From the recognition results, we can see that our algorithm can efficiently recognize this kind of targets.

A novel and effective immune multi-objective clustering algorithm (IMCA) is presented in this study. Two conflicting and complementary objectives, called compactness and connectedness of clusters, are employed as optimization targets. Besides, adaptive ranks clone, variable length chromosome crossover operation and k-nearest neighboring list based diversity holding strategies are featured by the algorithm. IMCA could automatically discover the right number of clusters with large probability. Seven complicated artificial data sets and two widely used synthetic aperture radar (SAR) imageries are used for test IMCA. Compared with FCM and VGA, IMCA has obtained good and encouraging clustering results. We believe that IMCA is an effective algorithm for solving these nine problems, which should deserve further research.

The problem of infrared targets detection in heavy clutter background is unsolved well. In this paper, we proposed a matched filter method based on tensor model, which generalizes the traditional matched filter, is used to distinguish between targets and clutter using multispectral infrared image. The tensor matched filter approach is adopted to take into account both spatial and spectral information simultaneously. The 3-mode product between tensor and filter matrix is given for matched filter. This method doesn't destroy the original spatial structure of multispectral images. By considering the spatial and spectral information for the filter, the probability of detection is improved. And an improved multi-way filter for multispectral tensor is performed before tensor matched filter. The experimental results of ROC curves indicate that our algorithm is effective to improve detection performance of targets in clutter background. And the multispectral images acquired at 8-9 μm , 9-10 μm , 10-11 μm , 11-12 μm , and 8-14 μm were tested in the experiment. We also show that this novel approach is extremely efficient in applications of interest by using tensor model for multispectral IR image processing.

Accurate co-registration of synthetic aperture radar (SAR) images is essential for SAR interferometric applications such as along track interferometry (ATI) and across track interferometry (XTI). In this paper, we present a non-linear and linear joint algorithm for the SAR image co-registration which could be used in highly rugged terrain. First, a slave image transformation is performed by a non-linear mapping which is based on free available SRTM (Shuttle Radar Topography Mission) DEM (Digital Elevation Models) and orbit parameters of both tracks to correct the distortions related to different view angles. Second, we calculate linear offsets between master image and transformed slave image with ordinary correlation based method to correct the offsets related to orbital errors. Our initial results show the effectiveness of such algorithm in co-registering critical SAR images.

The thematic remotely sensed information extraction is always one of puzzling nuts which the remote sensing science faces, so many remote sensing scientists devotes diligently to this domain research. The methods of thematic information extraction include two kinds of the visual interpretation and the computer interpretation, the developing direction of which is intellectualization and comprehensive modularization. The paper tries to develop the intelligent extraction method of feature space transformation for the deciduous forest thematic information extraction in Changping district of Beijing city. The whole Chinese-Brazil resources satellite images received in 2005 are used to extract the deciduous forest coverage area by feature space transformation method and linear spectral decomposing method, and the result from remote sensing is similar to woodland resource census data by Chinese forestry bureau in 2004.

Remote sensing image based on the complexity of the background features, has a wealth of spatial information, how to extract huge amounts of data in the region of interest is a serious problem. Image segmentation refers to certain provisions in accordance with the characteristics of the image into different regions, and it is the key of remote sensing image recognition and information extraction. Reasonably fast image segmentation algorithm is the base of image processing; traditional segmentation methods have a lot of the limitations. Traditional threshold segmentation method in essence is an ergodic process, the low efficiency impacts on its application. The ant colony algorithm is a populationbased evolutionary algorithm heuristic biomimetic, since proposed, it has been successfully applied to the TSP, job-shop scheduling problem, network routing problem, vehicle routing problem, as well as other cluster analysis. Ant colony optimization algorithm is a fast heuristic optimization algorithm, easily integrates with other methods, and it is robust. Improved ant colony algorithm can greatly enhance the speed of image segmentation, while reducing the noise on the image. The research background of this paper is land cover classification experiments according to the SPOT images of Qinling area. The image segmentation based on ant colony algorithm is carried out and compared with traditional methods. Experimental results show that improved the ant colony algorithm can quickly and accurately segment target, and it is an effective method of image segmentation, it also has laid a good foundation of image classification for the follow-up work.

Satellite imagery provides a cost-effective way to retrieve the cyanbacteria bloom dynamics, which is useful to early warning of the blooms. However, temporal variations in sun-target-satellite geometry and atmosphere may generate inconsistencies in multi-temporal images. To explore to what extent temporal influences could affect the retrieved results, we applied the single band and the band ratio approaches to retrieve cyanobacteria bloom in Lake Taihu of China. We used the Moderate Resolution Imaging Spectroradiometer (MODIS) products in the cases with and without correction for sun-target-satellite geometry and atmospheric effects for the whole year 2006. In addition, we made use of MODIS data including aerosol optical thickness (AOT), solar zenith angle and sensor zenith angle, all of which are indicators of the temporal influences. We then analyzed the relationships of retrieval differences with the three indicators to evaluate the temporal influences quantitatively. Our results showed that both AOT and solar zenith angle had a positive correlation with the retrieval of cyanobacteria bloom. Although it is yet under investigation if this relationship could hold on for other cases, here we emphasized that for reliable monitoring the dynamics of bloom, it should be careful to apply the approaches using satellite data without radiometric correction.

In this paper, we study the lossless and lossy compression methods for MODIS Level 0 data. Firstly, in order to decrease the compression time we use the simple image partitioning algorithm for MODIS image before compression. The experimental results show that by using the partitioning algorithm the compression time can be decreased about 25%. Then we present some lossless and lossy compression methods for MODIS level 0data. For lossless compression the experiments show that using the interband predictor helps to improve the compression rate because the redundancy among bands is reduced. For lossy compression we utilize some wavelet based methods to compress the denoised images. Finally we present the lossless and lossy compression results for MODIS image data by the methods.

For the fusion of urban multi-spectral images, three extracted characteristics are merged by a multi-characteristics fusion rule. The fusion result performs well in both spectral information and spatial information.

Spectral information and spatial information of high spatial resolution remotely sensed data are useful for natural disasters remote sensing application. At present, spatial resolution and spectral information of high spatial resolution remotely sensed images are contradictory factors. Image fusion methods can balance the contradictions between spatial resolution and spectral information. In this paper, three image fusion methods such as SFIM, WT and MBT have been employed to fuse either individual band multispectral images with panchromatic of FORMOSAT-2 in disaster area of Wenchuan earthquake, Sichuan province, China. Qualitative evaluation and quantitative assessment confirm that MBT can gain more texture information from panchromatic image and MBT is a good choice when image fusion is used in accurate ploting or locating the heavy disaster area and analyzing the loss of different disaster-bearing body using visual interpretation. While WT can preserve more spectral characteristic of FORMOSAT-2 multispectral images and WT will be a suitable image fusion method for disaster remote sensing applications when more spectral information is need.

Spectral unmixing (SU) is a hot topic in remote sensing image interpretation, where the linear mixing model (LMM) is discussed widely for its validity and simplicity [1]. SU often includes two facts as follows: 1) endmembers extraction; 2) abundances estimation. Mathematically, in the SU model, the collections, the endmember signatures, and the abundances are nonnegative [1]. Therefore, nonnegative matrix factorization (NMF) has a great potential to solve SU, especially for LMM [2]. In fact, NMF (or NMF like) algorithms have been widely discussed in SU, such as NMF based on minimum volume constraint (NMF-MVC) [1], NMF based on minimum distance constraint (NMF-MDC) [3], and so on. These methods have advantages and disadvantages, respectively. In light of that the abundances are often sparse and sparse NMF tends to result more determinate factors, NMF with sparseness constraint has attracted more and more attentions [4-6].To solve SU using sparse NMF practically, one problem should be addressed firstly, that is how to select the functions to measure the sparseness feature. Since the abundance suffers from sum-to-one constraint physically, the widely used measure based on L1 norm constraint may be degenerate [7, 8]. As the smoothed L0 norm of the signals can reflect the sparseness intuitively and it is easy to be optimized, we focus on NMF with smoothed L0 norm constraint (NMF-SL0) in this work [9]. The rest of this paper is organized as follows. In Section II, typical SU and NMF models are presented. Section III describes the L0-based sparse NMF for solving SU, together with the gradient based optimization algorithm NMF-SL0. Simulations using synthetic mixtures and real hyperspectral images are presented in Section IV. Finally, conclusions are summarized in Section V.

Doppler weather radars are capable of providing high quality wind data at a high spatial and temporal resolution. However, operational application of Doppler velocity data from weather radars is hampered by the infamous limitation of the velocity ambiguity. This paper reviews the cause of velocity folding and presents the unfolding method recently implemented for the CINRAD systems. A simple interactive method for velocity data, which corrects de-aliasing errors, has been developed and tested. It is concluded that the algorithm is very efficient and produces high quality velocity data.

Image fusion information extracted from multiple images which is more accurate and reliable than that from just a single image. Since various images contain different information aspects of the measured parts, and comprehensive information can be obtained by integrating them together. Image fusion is a main branch of the application of data fusion technology. At present, it was widely used in computer vision technology, remote sensing, robot vision, medical image processing and military field. This paper mainly presents image fusion's contents, research methods, and the status quo at home and abroad, and analyzes the development trend.

The point objectives in the hyper-spectrum image are difficult to be identified by the geometrical figure. It is needed to reduce the dimension of the spectrum data in order to eliminate the Information superabundance of the multi-dimension hyper-spectrum image data. We develop the distributed rules of the point objectives scatter plot in the low-dimension space and confirm that the point objectives End Member are mainly distributed in the lesser confidence interval while with the higher confidence coefficient. Finally we put forward to eliminate the non-point objectives and the noise End Member going beyond the threshold so as to ensure the result of the characteristic clustering is effective. Based on the scatter plot analysis, we find the new method to extract the spectrum characteristics by which we combine the mathematics analytical models, statistical computing and the distinguishing effects tests. At the same time we establish the model of spectrum character distinguish. According to the basic characteristics of the spectrum reflection features in the green vegetation we confirm two kinds of characteristic bands, setting up the training type, and one-dimension vector is formed after sampling by linearity combination. Through the practical application, we find the rather perfect spectral classification characteristics and the discriminant function for both the point objectives and background.

The classification of remote sensing images is a key issue and hot topic in remote sensing image processing domain. Considering that the classification result of classical principle component analysis (PCA) is not satisfying when the spectra of different ground objects are related, a new classification method based on sparse component analysis (SCA) is presented. The proposed method utilizes the sparse characteristic to extract the source signals, and does not demand the sources be independent. The experimental result of TM image shows that compared to the PCA method the overall classification precision of the SCA method enhances approximately 15%, which indicates that the classification result of the SCA method is more reliable and more accurate.

We describe a new real-time multispectral dynamic scene simulation system in this paper. Based on the spectral signatures, we model the multispectral flying targets for scene simulation and consider the atmospheric attenuation effect to enhance the scene reality. A key-frame based pre-calculation rendering algorithm is also present to accelerate the scene simulation speed. We also realize the multi-resolution multispectral backgrounds texture generation and real-time loading. In our simulation system, users can interactively real-time change the scene wavebands and viewpoints to observe the flying targets. Users can also play back the videos of the dynamic processes for further analysis after the simulation. The average dynamic rendering speed of our scene simulation is larger than 80fps and the experiments show the potential of our multispectral dynamic scene simulation system.

This paper proposes a fast approach to spectral image segmentation. In the algorithm, two popular techniques are extended and applied to spectral color images: the mean-shift filtering and the kernel-based clustering. We claim that segmentation should be completed under illuminant F11 rather than directly using the original spectral reflectance, because such illumination can reduce data variability and expedite the following filtering. The modes obtained in the mean-shift filtering represent the local features of spectral images, and will be applied to segmentation in place of pixels. Since the modes are generally small in number, the eigendecomposition of kernel matrices, the crucial step in the kernelbased clustering, becomes much easier. The combination of these two techniques can efficiently enhance the performance of segmentation. Experiments show that the proposed segmentation method is feasible and very promising for spectral color images.