Sea ice conditions are so heterogeneous, and the differences between the different ice types are less varied than that of land targets, so only using polarimetric or textural features would lead to misclassification of polarimetric synthetic aperture radar (PolSAR) data of sea ice. To support the identification of different ice types, the fusion of textural and polarimetric features would be a good solution. Simple discrimination analysis is used to rationalize a preferred features subset. Some features are analyzed, which include entropy H/alphaα/anisotropyA and three kinds of texture statistics (entropy, contrast, and correlation), in the C- and L-band polarimetric mode. After that, a multiobjective fuzzy decision model is proposed for supervised PolSAR data classification of sea ice, and the targets are categorized according to the principle of maximum membership grade. In consideration of the interference of the correlation among features, the model is based on Mahalanobis distance in which the covariances between the selected heterogeneous features could restrain the interference among redundant features. In the end, the effectiveness of the algorithm for PolSAR image classification of sea ice is demonstrated through the analysis of some experimental results.
Oil spills could occur in many conditions, which results in pollution of the natural resources, marine environment and economic health of the area. Whenever we need to identify oil spill, confirm the location or get the shape and acreage of oil spill, we have to get the edge information of oil slick images firstly. Hyperspectral remote sensing imaging is now widely used to detect oil spill. Active Contour Models (ACMs) is a widely used image segmentation method that utilizes the geometric information of objects within images. Region based models are less sensitive to noise and give good performance for images with weak edges or without edges. One of the popular Region based ACMs, active contours without edges Models, is implemented by Chan-Vese. The model has the property of global segmentation to segment all the objects within an image irrespective of the initial contour. In this paper, we propose an improved CV model, which can perform well in the oil spill hyper-spectral image segmentation. The energy function embeds spectral and spatial information, introduces the vector edge stopping function, and constructs a novel length term. Results of the improved model on airborne hyperspectral oil spill images show that it improves the ability of distinguishing between oil spills and sea water, as well as the capability of noise reduction.
As the hyperspectral image combines spacial information with spectral information, the spectroscopic data can describe the characteristics of surface feature more accurately, and provide possibilities for classification and quantitative calculation for the surface features. Now, the unmixing technology for mixed pixel has become a hot topic in this field. The technology for pixel unmixing contains two main directions. The first one is based on linear mixing model. This model assumes that the pixel is formed by endmembers according to linear relationship. The methods based on this model are easy to be implemented. But the ideal model can’t describe the mixed relation of the surface features accurately. So the accuracy of abundance estimation can’t be guaranteed. The second one is based on non-linear model. This method could get good analytical results, but they are mainly established for particular surface features and difficult for implementation. This paper was mainly aimed at the research of abundance estimation. A simplified Hapke model is proposed to be applied to actual hyperspectral image of oil spilling, so as to obtain the estimation of oil thickness. The Hapke model could transform the non-linear relationship to linear relationship for hyperspetral data. The spectral reflectances of non-linear relationship are transformed to spectral albedo satisfying linear relationship, without changing the abundance. At last, this model is applied to actual hyperspectral image of oil spilling, achieving estimation for oil thickness.
Hyperspectral remote sensing has been widely used in more and more fields nowadays, such as the oil spill analysis and chlorophyll estimation in green plants. To decompose the mixed pixels people always turns to the traditional method of Least Squares Method now. But its main drawback is that it involves a large amount of matrix operations, especially regarding to the huge dimension of hyperspectral images. So it will take much time. Motivated by this, in this paper we have developed a new model of endmember abundance estimate which is referred to as Spectral Characteristic Based Abundance Estimation Model (SCBAEM). The model is based on the fitted curve in which spectral characteristic were considered. To establish the model, Spectral Angle Mapper (SAM) and Spectral Information Divergence (SID) were utilized between endmember and mixed pixels. The main contributions of the paper are summarized as follows: Firstly, we build the model by calculating normalized SAM (NSAM) and normalized SID (NSID). Secondly, to test and verify the accuracy of the model, oil slick experiment is carried out. Finally, we further conduct its application in the real hyperspectral oil spill images which is from Peng-lai 19-3C platform. The results of simulation experiments and real hyperspectral image demonstrate that the proposed model could achieve the efficiency of LSM. At the same time, the time cost can be reduced greatly. So it can satisfy the real-time need.
In this paper, we address the problem of extracting the edge of power lines from aerial images, which is a critical step
for the identification of the components equipped on power lines and for the diagnosis of the broken-strands on power
lines. As for the problem, a novel idea is proposed based on the fact that the textural differentiation of Dissimilarity
distributions between power line and background in a square window can be represented by some points of local peaks
on the statistical curves such that the edge of power lines can be detected. Based on the novel idea, an algorithm named
Square Window wIth Four symmeTrical axiSes (SWIFTS) is developed. Experiments and analyses demonstrate that the
SWIFTS algorithm has better performance than other classical method in terms of extracting the edge of power lines,
preserving the bent information of the edge at the point of broken-strand.
Spilled oil is one of the most serious marine environment disasters, which damaged ecological environment seriously
with long-term and large-scale impact. Based on the experiment and research in the Canadian Centre of Environmental
Technology, an experiment is taken to detect the underwater suspended oil-spills by Laser-induced fluorescence. It
quantizes the conditions that Laser-induced fluorescence can be used to detect underwater oil, and makes a solid theory
foundation for the system design of underwater oil detection by Laser-induced fluorescence. This environmental solves a
key problem for underwater oil detection by Laser-induced fluorescence.
The extraction of power lines from aerial images and the distinguishing of them are of great significance to locate the
points of the faults on power lines and to diagnose the components equipped on power lines. To solve the two problems,
an algorithm composed of three steps is proposed as follows. Firstly, the edges of power lines are extracted by SWIFTS
algorithm1, which is developed on the basis of the differentiation of Dissimilarity between power line and background.
Secondly, to detect power lines, Hough transform (HT) is employed by taking advantage of its insensitive to noise.
Lastly, a clustering algorithm based on Nearest Neighborhood (NN) is adopted to distinguish power lines. Experiments
and analysis demonstrate that the proposed algorithm could effectively extract and distinguish power lines from aerial
images.
Remote sensing data cover large areas and can be acquired in a regular repeatable manner. Automatic land-cover
classification in satellite images is an important topic and has applied in remote sensing widely. In this paper, we
consider Landsat5 Thematic Mapper (TM) data of the Qinghai-Tibet highway of 1986 and 1994 to analyze the changes
of land-cover. Statistics and artificial intelligence method are combined to improve the classification precision. And the
classification result can provide quantitative data for road environment issue, road location selection, and landscape
design.
In this paper, Principal Component Analysis (PCA) is applied to character the main information of TM land-cove
image. Then two neural network models are used to classify the TM image: Back-Propagation Neural Network (BPNN)
and Self-organizing feature map (SOFM) neural network. BP neural network is widely used. Contingency matrix is used
to evaluate the classification precision. By comparing classification accuracy and Kappa quotient, conclusion is drawn
that the classification accuracy of SOFM is higher than BP and MLC and the classification ability of BP is not as good as
MLC. Overall accuracy of SOFM is 94.0%, Kappa is 0.9114, and overall accuracy is 14.9% and 9.8% higher than BP
and MLC. So SOFM is used to classify image of 1986. In the end the land-cover changes of two year are analyzed.
This article proposed a method for the recognition of the sea ice in a SAR image. It's based on the Artificial Neural Network (ANN). We use a BP neural network model incorporating with image texture features extracted by Gray Level Co-occurrence Matrix from the SAR image. The BP neural network fed with the feature vector of SAR image presents the analysis of texture features and outputs the estimation results of the sea ice. The BP neural network is trained using sample data set to the neural network. And then the BP neural network trained is tested to recognize sea ice in a SAR image waiting for the classification . The results of tests show that the BP network model for the sea ice recognition in SAR image is feasible. The BP network output shows the recognition accuracy of the model for the sea ice recognition in SAR image can be 86%. Among the 4 directions for computing the texture features, the most valuable direction is 0°, the next is 90°, and then is 135°. The two most distinguishing texture features are inertia moment and entropy.
A Radial Basis Function Neural Network (RBFNN) Model is investigated for the detection of spilled oil based on the texture analysis of SAR imagery. In this paper, to take the advantage of the abundant texture information of SAR imagery, the texture features are extracted by both wavelet transform and the Gray Level Co-occurrence matrix. The
RBFNN Model is fed with a vector of these texture features. The RBFNN Model is trained and tested by the sample data set of the feature vectors. Finally, a SAR image is classified by this model. The classification results of a spilled oil SAR image show that the classification accuracy for oil spill is 86.2 by the RBFNN Model using both wavelet texture and gray texture, while the classification accuracy for oil spill is 78.0 by same RBFNN Model using only wavelet texture as the input of this RBFNN model. The model using both wavelet transform and the Gray Level Co-occurrence matrix is more effective than that only using wavelet texture. Furthermore, it keeps the complicated proximity and has a good performance of classification.
Oil spills are very serious marine pollution in many countries. In order to detect and identify the oil-spilled on the sea by remote sensor, scientists have to conduct a research work on the remote sensing image. As to the detection of oil spills on the sea, edge detection is an important technology in image processing. There are many algorithms of edge detection developed for image processing. These edge detection algorithms always have their own advantages and disadvantages in the image processing. Based on the primary requirements of edge detection of the oil spills’ image on the sea, computation time and detection accuracy, we developed a fusion model. The model employed a BP neural net to fuse the detection results of simple operators. The reason we selected BP neural net as the fusion technology is that the relation between simple operators’ result of edge gray level and the image’s true edge gray level is nonlinear, while BP neural net is good at solving the nonlinear identification problem. Therefore in this paper we trained a BP neural net by some oil spill images, then applied the BP fusion model on the edge detection of other oil spill images and obtained a good result. In this paper the detection result of some gradient operators and Laplacian operator are also compared with the result of BP fusion model to analysis the fusion effect. At last the paper pointed out that the fusion model has higher accuracy and higher speed in the processing oil spill image’s edge detection.
In this paper an artificial neural network (ANN) approach, which is based on flexible nonlinear models for a very broad class of transfer functions, is applied for multi-spectral data analysis and modeling of airborne laser fluorosensor in order to differentiate between classes of oil on water surface. We use three types of algorithm: Perceptron Network, Back-Propagation (B-P) Network and Self-Organizing feature Maps (SOM) Network. Using the data in form of 64-channel spectra as inputs, the ANN presents the analysis and estimation results of the oil type on the basis of the type of background materials as outputs. The ANN is trained and tested using sample data set to the network. The results of the above 3 types of network are compared in this paper. It is proved that the training has developed a network that not only fits the training data, but also fits real-world data that the network will process operationally. The ANN model would play a significant role in the ocean oil-spill identification in the future.
In order to detect and identify oil-spilled on the sea by Airborne Laser-Induced Fluorescence, a fuzzy model and algorithm are put forward in this paper. The target to be detected on the sea may be one of the following: seawater, crude oil, diesel, lubricating oil, dirty water, sand, etc. The primary requirement for airborne sensors is to identify, in real-time, the substances targeted by the laser beam. There have been several algorithms developed for the detection of oil spilled on the sea by Airborne Laser-Induced Fluorescence, for example, the Pearson Correlation Coefficient method. The reason that we have decided to research the fuzzy model for the identification of oils spilled on the sea, is that there are some uncertainties and unknown differences between the “live” spectrum, and the substances targeted by the laser beam. The fuzzy algorithm presented in this paper is based on a fuzzy closeness matrix. All values in the matrix are calculated from the spectrum of a target and the spectra of the above mentioned “pure” substances.
This paper outlines the fuzzy model for the identification of the spilled oils, and makes a comparison with the Pearson Correlation Coefficient method in an effort to increase the level of confidence in the identification results and reduce the computational time. The results of ground tests using known targets show an increased confidence with the identification results using the Fuzzy Model when compared to the results of the Pearson Correlation Coefficient Algorithm.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.