Looking toward 2000, the ways by which commercial satellite imagery and imagery products are managed by the various remote sensing companies will be dictated by financial considerations, not technical feasibility. In the convergence of technologies that will shape the commercial companies in 2000, the most influential will likely be electronic commerce via the Internet. This paper discusses the character of these combined forces and speculates on how the industry might respond.

The reference data consists of two or more central- projection images of a three-dimensional distribution of object points, i.e., a 3D scene. The positions and orientations of the cameras which generated the reference images are unknown, as are the coordinates of all the object points. We derive and demonstrate invariant methods for synthesizing nadir views of the object points, i.e., 2D maps of the 3D scene. The techniques we demonstrate depart from standard methods of resection and intersection to recover the camera geometry and reconstruct object points from the reference images, followed by back-projection to create the nadir view. Our approach will be to perform the image measurements and computations required to estimate the image invariant relationships linking the reference images to one another and to the nadir view. The empirically estimated invariant relationships can thereafter be used to transfer conjugate points from the reference images to their synthesized conjugates in the nadir view. Computation of the object model -- the digital elevation model (DEM) -- is not required in this approach. The method also differs from interpolation in that the 3D structure of the scene is preserved, including the effects of partial occlusion. Algorithms are validated, initially with synthetic CAD models and subsequently with real data consisting of uncontrolled aerial imagery and maps with occasional missing or inaccurately delineated features.

Development of vector feature extraction techniques combined with large amounts of digital map data can create a rich database for intelligence activities. Our ability to use this data depends on understanding the relationships between different features and different representations of the same feature. Linking is the process of determining that two features in different layers represent the same object. Links between features in two or more layers can be used for: co-registering vector data from multiple sources, automatic revisions of linear data, feature attribute inheritance, or network transversal problems. One sample application uses linking for generating a database with features extracted from stereo imagery and attributes from existing DMA data sources. Time and personnel are limited resources. Therefore, the linking process needs to be automated. Two data sets are used -- pre-existing road data at 1:250 K scale (Set A) and stereo imagery used for extracting 1:50 K roads (Set B). Set A is richly attributed. We want the spatial accuracy of the roads from Set B and the attributes of the roads from Set A. Linking can match the two sets of roads. A new combined data set is created in several stages. Roads where a one-to-one linkage exists between A and B use the spatial data from Set B and the attributes from Set A. Roads that are unique in either set are added. Attributes are retained for the roads from Set A. This paper discusses a new technique for automatic feature linking developed at GDE Systems Inc. and demonstrated in a prototype. The prototype uses the characteristics of a linear vector feature to identify the same features from different sources and from different scales.

The exploitation of remote sensing data across multiple applications is greatly facilitated by a platform for delivering the information that can be readily tailored for the requirements of specific problems, while providing a standard infrastructure for storage, metadata management, dissemination, and integration with desktop tools and other multimedia data relevant to a problem. We have developed such a platform for data generated by remote sensing instruments, including the TRW Imaging Spectrometer (TRWIS) family. Based on TRW's InfoWeb^TM digital multimedia archive architecture, it provides an object-relational, intranet environment for finding and manipulating hyperspectral data and derived products. in this paper, we discuss the platform architecture and strategies for data management and integration. We present lessons learned in applying this technology that are illustrative for the development of remote sensing delivery systems.

TRW, under a Small Satellite Technology Initiative (SSTI) contract, is building the Lewis satellite. The principal sensor on Lewis is a hyperspectral imaging spectrometer. Part of the SSTI mission is to establish the commercial and educational utility of this data and the hyperspectral data already being acquired on airborne platforms. Essential requirements are rapid availability (after data acquisition) and easy accessibility to a catalog of images and imagery products. Each image is approximately 256 by 512 pixels with 384 bands of data acquired at each pixel. For some applications, some users will want the entire data sets; in other cases partial data sets (e.g. three band images) will be all that a user can handle or need for a given application. In order to make the most effective use of this new imagery and justify the cost of collecting it, we must find ways to make the information it contains more readily accessible to an ever broadening community of potential users. Tools are needed to store, access, and communicate the data more efficiently, to place it in context, and to derive both qualitative and quantitative information from it. A variety of information products which address the specific needs of particular user communities will be derived from the imagery. The data is unique in its ability to provide high spatial and spectral resolution simultaneously, and shows great promise in both military and civilian applications. A data management and analysis system has been built at TRW. This development has been prompted by the business opportunities, by the series of instruments built here and by the availability of data from other instruments. The products of the processing system have been shown to prospective customers in the U.S. and abroad. The system has been used to process data produced by TRW sensors and other instruments. This paper provides an overview of the TRW hyperspectral collection, data handling and exploitation capability.

An experimental GIS for the Island of Sicily was developed, aiming at the classification of some traits of both the marine and terrestrial environment in coastal zones. The GIS integrates data generated by remote sensors and conventional cartographic data. Parameters describing the marine environment, at the regional scale, were derived from time series of composite chlorophyll-like pigment concentration images generated by the Coastal Zone Color Scanner (CZCS) in the period 1979 - 1985, and of composite sea surface temperature images generated by the Advanced Very High Resolution Radiometer (AVHRR) in the period 1982 - 1985. Other auxiliary data, such as bathymetry and maps of hydrological basins, rivers, elevation, and reservoirs were selected to complement the low-resolution satellite data set and introduced in the regional GIS. At the local scale (Simeto river basin and Gulf of Catania), parameters describing essential coastal water constituents and vegetation were derived from TM data (1985 and 1994). The ancillary data set integrated with the high-resolution images includes information about physical and biological components of the local coastal environment (i.e. geographic, geological, geomorphologic, sedimentary and vegetation outlines). Examples of the synergistic use of such data, at various space/time scales, include assessing the impact of runoff from hydrological basins on the marine ecosystem, sediment transport and coastal evolution, vegetation cover.

Future multispectral and hyperspectral remote sensing systems and image archives will benefit from effective, high-fidelity image compression techniques. In evaluating the effects of compression upon the data, one must not only consider the qualitative and quantitative effects upon the images themselves, but also upon the end user products that are derived from the imagery through the application of environmental retrieval algorithms. At The Aerospace Corporation, we have developed a fast algorithm for image compression techniques known as the modulated lapped transform (MLT). This compression algorithm obviates many of the artifacts that are introduced by some of the standard compression techniques. One example of compression artifacting is the blocking errors from discrete cosine transformation (DCT) based algorithms, which include the JPEG compression scheme. The Aerospace MLT technique is a hybrid of the wavelet and DCT techniques. It employs our patented split-radix approach, which is the fastest DCT algorithm known today. In this paper, we compare Aerospace MLT to JPEG, using cloud imagery and Earth surface scene classification. We also discuss the availability of a cost- effective VLSI hardware implementation of the Aerospace compression algorithm. The modulated lapped transform employs a peano scan with a split-radix approach to avoid blockiness artifacts. It has excellent resistance to errors, and it is amenable to fast processing using a 1-D hardware architecture to process a 2-D image. This technique encapsulates the favorable aspects of the wavelet transforms and produces images which, when compressed 10:1 and decompressed, compare very favorably (using error statistics, classification accuracy and visual quality metrics) to the original uncompressed image.

The conversion of raw imagery to information products tailored to specific user needs will become increasingly important in order to maximize the utility of the data and deal with the large amount of imagery that satellite systems will be capable of collecting. This conversion will need to be largely automated if the resulting products are to be cost effective. By making available the third, or height, dimension, collection of overlapping imagery of the same region helps both to broaden the class of information products that can be produced and to increase the reliability of automated processes for information extraction. This paper describes processes for extraction of height information and enumerates some of the products which such information enables.

Syseca and IGN are working on various steps in the ongoing march from digital photogrammetry to the semi-automation and ultimately the full automation of data manipulation, i.e., capture and analysis. The immediate goals are to reduce the production costs and the data availability delays. Within this context, we have tackle the distinctive problem of 'automated road network extraction.' The methodology adopted is to first study semi-automatic solutions which probably increase the global efficiency of human operators in topographic data capture; in a second step, automatic solutions are designed based upon the gained experience. We report on different (semi-)automatic solutions for the road following algorithm. One key aspect of our method is to have the stages of 'detection' and 'geometric recovery' cooperate together while remaining distinct. 'Detection' is based on a local (texture) analysis of the image, while 'geometric recovery' is concerned with the extraction of 'road objects' for both monocular and stereo information. 'Detection' is a low-level visual process, 'reasoning' directly at the level of image intensities, while the mid-level visual process, 'geometric recovery', uses contextual knowledge about roads, both generic, e.g. parallelism of borders, and specific, e.g. using previously extracted road segments and disparities. We then pursue our 'march' by reporting on steps we are exploring toward full automation. We have in particular made attempts at tackling the automation of the initialization step to start searching in a valid direction.

Spatial evolutions of the anthropized ecosystems and the progressive transformation of spaces in the course of time emerge more and more as a special interest issue in research about the environment. This evolution can present a large preoccupation in space accommodation and environmental domains, and it gives rise to a considerable problem in terms of prospective. How will be the conditions of a region area, between now and 15, 30, or 50 years? In fact, the time consists of hierarchical events and can produce transformations upon a terrain landscape as emergence, disappearing, union of spatial entities. These transformations are called temporal phenomena. We propose to predict the forestry evolution in the forthcoming years on an experimental area which reveals these spatial transformations. For these purposes, we have developed a specific spatio-temporal prediction approach. The idea we present here takes a first step in attacking this problematic, it turns out very interesting results in this domain. We describe in this paper a method for analysis and prediction of terrain landscape for an established date. This method is founded on n geographic maps representing the terrain conditions for distinct years. The basic idea is to employ the observation of the temporal phenomena evolution. In fact, results of this observation represent the evolution of each region area on maps in the course of time. The evolution modeling of the regions is obtained with the help of a sequence of aerial photographies compared through different dates. It allows the geographer interested in environmental prospective problems to get type cartographical documents showing the future conditions of a landscape. This method makes use of vectorial geographic data and it achieves a prediction by means of different comparisons between attributes of regions such as the surface, center and distance between regions. The final shapes and positions of the regions are determined by combining the results stemming from applications of a linear regression method and from mathematic morphology in vectorial domain. The implemented approach model the evolution of the forest in a region of the south of France by using maps for the years 1942, 1962, and 1993. We used this method to study a region located in the Ariege mountains called 'Soulave' to describe the evolution of its landscape for the years 2000, 2005, 2010, 2015, and 2020. The experimental tests have shown promising results.

Several methods have been proposed for the extraction of latent information from multispectral remotely sensed scenes based on the definition of indices and rotational transformations. A common drawback of these techniques is that they are ultimately based only on statistical relationships among pixel values rather than on physical characteristics of the scenes. Linear pixel unmixing is an alternative method which assumes that the pixel signal is the linear combination of some basic spectral components the fractions of which can be retrieved with good approximation. The method is straightforward and produces results which can be easily interpreted, but presents the problem of the identification of suitable end-members, which generally requires some external knowledge. In order to overcome this problem, in the present research a statistical method is developed for the automatic identification of end-members. This methodology is composed by several steps, that are describe and then applied to a case study with a Landsat 5 TM scene from Central Ethiopia (Africa). The results, evaluated in comparison with those of a more usual principal component transformation, indicate the good performance of the new procedure.

In this paper, the effectiveness of multifrequency polarimetric SAR data for archaeological studies is assessed. We propose two detail-preserving segmentation algorithms for multiband images and evaluate their performance using real multi-frequency multi-polarization SAR complex data. Adaptive neighborhood structures are selected for modeling the polarimetric complex amplitudes and the region labels, and for achieving detail- preservation. Experimental results obtained from multi-band and multi-polarization SIR-C data, selected for archaeological applications studies, show that the novel schemes produce significant visual improvements for detail preservation, and exhibit equivalent or higher classification performance with respect to the classical classification schemes.

Nowadays the study of pre/protohistoric communities is performed by archaeologists taking account either the information gathered from field work surveys and excavations or the characteristics of the surrounding landscape, on the basis of a multi-disciplinary approach. The region considered in the present research is located in central Italy, northwest of Rome, between the Bolsena Lake and the Tyrrhenian Sea. This area exhibits two different landscape typologies represented by the Albegna and Fiora rivers catchments where more than 300 archaeological sites have been detected. The analysis has concerned only 112 sites corresponding to graveyards and settlements, whose respective geographical coordinates have allowed us to gather peculiar environmental parameters. These ones have been obtained from different data sets: the six TM bands of a LANDSAT subscene; a digital elevation model (DEM), with derived thematic maps; hydrological and hydrogeological features extracted from suitable scale maps. The hydrological data are represented by streams branches, drainage divides, main spring sites, lakes and sea cost lines, piezometric contours (to assess water table depth), hydrogeological unit identifications (to provide lithological and permeability information). The obtained parameters are: morphometric (i.e. elevation, slope, aspect, local relief and morphometric code from the DEM), spectral (i.e. the TM values of the Landsat data) and hydrological (minimum distance from water resources, water table depth). They have been processed by applying a hierarchical clustering technique separately to the morphometric and spectral parameters, achieving as a result the partitioning of the sites into meaningful classes, characterized by different morphology, land cover and physical status. The obtained groups of sites have been examined with respect to their typological and chronological descriptions.

Several attempts have been made to merge Landsat TM multispectral data with high spatial resolution panchromatic SPOT images. In this work a multiresolution approach based on a generalized Laplacian pyramid with p:q (i.e., rational) scale factor is proposed to merge image data of any resolution and represent them at any scale. The resulting bands capture multispectral characteristics with an enhanced spatial resolution, thereby expediting visual analysis and contextual interpretation of the environment according to archaeological issues. Objective and subjective evaluations show the effectiveness of the proposed method.

The Getty Conservation Institute is exploring the feasibility of using remote sensing associated with a geographic database management system (GDBMS) in order to provide archaeological and historic site managers with sound evaluations of the tools available for site and information management. The World Heritage Site of Chaco Canyon, New Mexico, a complex of archeological sites dating to the 10th to the 13th centuries AD, was selected as a test site. Information from excavations conducted there since the 1930s, and a range of documentation generated by the National Park Service was gathered. NASA's John C. Stennis Space Center contributed multispectral data of the area, and the Jet Propulsion Laboratory contributed data from ATLAS (airborne terrestrial applications sensor) and CAMS (calibrated airborne multispectral scanner) scanners. Initial findings show that while 'automatic monitoring systems' will probably never be a reality, with careful comparisons of historic and modern photographs, and performing digital analysis of remotely sensed data, excellent results are possible.

Lidars have already been used as an effective tool for the monitoring of marine environment and vegetation. In this paper the data relative to the first fluorescence lidar experiments on the facades of historical buildings are presented and discussed. The lidar fluorescence spectra presented in this paper were detected with a high spectral resolution fluorescence lidar on the Baptistery of Parma and mainly refer to the coatings of Ammonitico Rosso Veronese, a calcareous stone widely employed as decorative material for its color and typical texture. These data were also compared to the lidar spectra of samples coming from different sites situated in the extraction areas historically known from the archives and from sites in an area exploited only quite recently. The results constitute a first step towards a completely non destructive spectral analysis directly on the surfaces of the historical buildings.

Fluorescence techniques have been extensively used for the detection of photoautotrophic organisms. In particular, fluorescence lidars have been successfully employed for the detection and identification of phytoplankton in sea water and, more recently, have been regarded as powerful tools for vegetation monitoring, especially as for the evaluation of forest decline. In this paper the fluorescence lidar technique is proposed as a new tool for the remote monitoring of photoautotrophic biodeteriogens on surfaces of historical monuments. The use of fluorescence lidars can remotely detect photoautotrophic organisms on monuments and, with a suitable spectral resolution, identify the pigments contained into. The advantages of the fluorescence lidar technique with respect to the traditional methods are manifold and can lead to a fast, extensive control of the stony cultural heritage at low cost. Laboratory and lidar experiments were carried out on different stony materials inoculated with microalgal and cyanobacterial species at different cell concentrations to investigate the potentialities of this technique. The experimental results presented in this paper include the remote detection of biodeteriogens on stony materials at a previsual growth stage by means of a fluorescence lidar and the identification of the fluorescence features of different pigments.

In this paper we discuss some observational results on the emission plume dynamics obtained by the airborne lidar during complex observations of the Far East cities. The experimental techniques were observed and are recorded in this paper.

It is well known that lidar technique of investigation of large scale aerosol pollutions has good outlook. However, difficulties of solution of lidar sounding equation and relationship between backscatter radiation and aerosol mass concentration exist. So one of the methods of increasing the accuracy of lidar measurements is discussed. It is based on the use of contact devices for aerosol measurement, which are operated simultaneously with lidar on board the plane.

Various mesoscale meteorological phenomena often determine the transportation, changes and spreading of the atmospheric pollutant over urban and rural areas with a developed industry. The corresponding investigations can be performed by combined use of remote (lidar) and conventional means. The present investigation aims at applying the accumulated experience for a determination of the air quality over large area (400 sq km) including residential districts of settlement in a vicinity of oil-refinery and for a determination of the mesoscale phenomena influence on the air quality of the resort in the coastal zone. We used a mobile aerosol lidar, a tethered balloon, pilot balloon measurements and point chemical analyses for the purposes of the mentioned investigation. The measurements were carried out from 14 April to 9 May 1992 in the Bourgas region.

This presentation, taking its inspiration from various theoretical contributions suggested by various proposals, and going beyond the usual explanations and sometimes rash forecasts that the impact of a new technology has on physical space, wishes to underline the most important aspects of the introduction of a new method of studying urban installations, both ancient and modern, which is offered by urban telesurveys. To this end it is proposed, by means of an analysis of a sample area obtained from tele- surveying data provided by an aerial survey, using an advanced system of electronic pictures shot by MIVIS AA500000 of the CNR-Progetto LARA, to reveal and elaborate various findings and thus highlight a quantity and quality of extremely useful data.

During the past year, Kestrel Corporation has designed and built a low cost Fourier transform visible hyperspectral imager (FTVHSI) for deployment in a light aircraft (Cessna TU-206). The instrument is an imaging spectrometer employing a Sagnac (triangle) interferometer, that operates over a range of 450 - 1050 nm with 256 spectral channels, and a 13 degree FOV with an 0.8 mrad pixel IFOV (450 spatial channels). To aid in the calibration of the instrument, calibration and downwelling signals are recorded with every frame. Installed with the optical instrument are attitude sensors and a scene camera. This auxiliary data allows us to place a hyperspectral slice to within less than 5 m of its true position (using selective availability 'on' and differential GPS). We have performed extensive testing and calibration studies, including data collection conducted synchronously with ground measurements at locations including a White Sands radiometric calibration site. This paper reports some of the calibration studies and their results.

Modern Earth sciences are beginning to examine interactions among the different terrestrial components at all temporal and spatial scales. Such a global perspective requires an integrated remote sensing strategy which uses instruments throughout the electromagnetic spectrum to collect data about the Earth's surface, oceans, and atmosphere over a range of selected scales. Studies of specific processes will require remotely sensed data at spatial, spectral, and temporal resolutions appropriate to the scale of the research.

With the goal of applying derivative spectral analysis to analyze high resolution, spectrally continuous remote sensing data, several smoothing and derivative computation algorithms have been reviewed and modified to develop a set of cross-platform spectral analysis tools. Emphasis was placed on exploring different smoothing and derivative algorithms to extract subtle spectral features from any continuous spectral data sets. With interactive selection of bandwidth and sampling interval (band separation), the algorithm can optimize noise reduction and better match the scale of spectral features of interest. Laboratory spectral data were used to test the performance of the implemented derivative analysis modules. An algorithm for detecting the absorption band positions was executed on synthetic spectra and a soybean fluorescence spectrum to demonstrate the usage of the implemented modules in extracting spectral features. Upon examination of the developed modules, issues related to the smoothing and the spectral deviation caused by the smoothing or derivative computation algorithms were also observed and discussed. The scaling effect resulting from the migration of band separations when using the finite approximation derivative algorithm was thoroughly inspected to understand the relationship between the scaling effect and noise removal.

The paper describes the first stage of an experiment, performed in the Fortore beneventano (Italy) mountain community, to test and enhance methods for mapping soil conditions from spectroradiometric measurements and hyperspectral images. This required as a pre-requisite the validation of MIVIS data characteristics and the correction of atmospheric and illumination effects resulting into the conversion of radiances to relative reflectance values. After radiometric rectification of the image data and the collection of a field/laboratory spectral library, linear spectral mixture modeling (SMA) was used to decompose image spectra into fractions of spectrally distinct mixing components. The resulting abundance estimates (fractions) then were analyzed to identify soil conditions as well as to obtain an improved measure of dry and green vegetation cover, which are considered important parameters for monitoring soil erosion processes and changes of vegetation cover density as indicators for decertification.

Study to develop a software methodology to geocode MIVIS hyperspectral images collected by the CNR LARA Project. Goal of the study is to integrate the airborne position and attitude system with the image data to obtain geocoded images at a medium-small scale (1:15000 - 1:10000).

This paper presents the results on the individuation of illegal waste sites, and relatively small objects in Northern Italy using airborne MIVIS data (multispectral infrared and visible imaging spectrometer; Deadalus AA5000). Information acquired by the sensors of the MIVIS hyperspectral system, developed by the National Research Council of Italy -- LARA (Airborne Laboratory for Environmental Research) project, presents three main characteristics: local scale study, possibility to plan the proper period based on the objectives of the study, high number of spectral bands with high spectral and geometrical resolution. Unlike to common mono and multispectral airborne and satellite remote sensing systems, MIVIS hyperspectral scanners provide such amount of data with high geometrical and spectral resolution that now it's no longer necessary to develop complex classification methodology in order to extract maximum information available. The problem is shifted to the selection of a relatively small number of bands, less correlated to each other, that emphasize and characterize even small targets. The flow of an operative program for the characterization, identification and classification of defined categories of objects is described. Requirements in data collection by airborne MIVIS data oriented to this target are defined. Flight planning, based on the targets like small or unusual features detection (e.g. roof skylight), is important to improve the usefulness of these sensors. The results obtained are encouraging. This instrument, supported by correct analysis techniques, may offer new interesting prospects in territorial analysis, small target detection and, consequently, environmental monitoring and management.

Environmental changes around the Oro Lake (Mali) which, with other lakes (Fati, Faguibine, Tele), represents an important hydrological resource of the Inland Delta of the Niger river, have been assessed by means of Landsat MSS imagery analysis. After data calibration and atmospheric correction, different color composites, band ratios and image differences were selected. Subsequently, the spectral features of the principal morphological units were analyzed using the satellite data, and were compared with the morphological characteristics detected and mapped in 1957 by Tricart from field work and aerial photographs. The morphological changes highlighted by satellite data therefore represent an interesting update of the old map in a region in which it is very difficult and expensive to perform ground surveys.

Most applications using remote sensing need the imagery to be georeferenced, and precision plays a fundamental role whenever cartographic or 'metric' aims are involved. Orthoimage generation techniques produce geometrically corrected images in which also the shift caused by reliefs is evaluated and removed. In addition to a digital terrain model (DTM) of the area, these methods require information on the orientation of the image and the attitude of the sensor-borne platform or, more in general, on the geometry of acquisition. In the case of satellite imagery, this information is not always available or accurate, in particular when sub-scenes have been extracted from a full frame. In this work a new georeferencing technique is presented which corrects the geometrical distortion due to the presence of reliefs. As auxiliary data, the proposed method needs only a DTM of the area and the identification of a set of ground control points, but information on the geometry of acquisition or on the characteristics of the sensor is not required. Two 3-D splines including height information are used, chosen so that they correctly match the ground control points and diverge far from them in linear manner. The proposed method appears to be more accurate compared to the use of polynomial best fit, because the high-frequency distortions due to height variations are accounted. for. Tests have been performed on an area including high-altitude variations.

The ancient Pompeii site is in the Sarno Valley, an area of about 400 km² in the South of Italy near Naples, that was utilized by man since old time (thousands of years ago). Actually the valley is under critical environmental conditions because of the relevant industrial development. ENEA is conducting various studies and research in the valley. ENEA is employing historical research, ground campaigns, cartography and up-to-date airborne multispectral remote sensing technologies to make a geographical information system. Airborne remote sensing technologies are very suitable for situations as that of the Sarno Valley. The paper describes the archaeological application of the research in progress as regarding the ancient site of Pompeii and its fluvial port.