Vision is a part of information system that converts visual information into knowledge structures. These structures drive the vision process, resolving ambiguity and uncertainty via feedback, and provide image understanding, which is an interpretation of visual information in terms of these knowledge models. It is hard to split the entire system apart, and vision mechanisms cannot be completely understood separately from informational processes related to knowledge and intelligence. Brain reduces informational and computational complexities, using implicit symbolic coding of features, hierarchical compression, and selective processing of visual information. Vision is a component of situation awareness, motion and planning systems. Foveal vision provides semantic analysis, recognizing objects in the scene. Peripheral vision guides fovea to salient objects and provides scene context. Biologically inspired Network-Symbolic representation, in which both systematic structural/logical methods and neural/statistical methods are parts of a single mechanism, converts visual information into relational Network-Symbolic structures, avoiding precise artificial computations of 3-D models. Network-Symbolic transformations derive more abstract structures that allows for invariant recognition of an object as exemplar of a class and for a reliable identification even if the object is occluded. Systems with such smart vision will be able to navigate in real environment and understand real-world situations.

In this paper, a new formulation for the parametric active contour model is presented. The new formulation is based on statistical pattern recognition theory. A hybrid of kernel density estimation and fuzzy logic is used to show that active contours can be thought of as a pattern recognition problem. The proposed approach is used in two different application domains, with different performance requirements, to demonstrate its effectiveness. First, the proposed approach is used for a magnetic resonance image segmentation problem to demonstrate the segmentation accuracy. Second, the contour is used in a target tracking experiment to show its tracking capabilities.

The research discussed in this paper is a continuation of the author's previous research published in SPIE's Visual Information Processing Proceedings of 2003 [1] entitled "Improving the Performance of Content-Based Image Retrieval Systems". The SPIE article discussed a new method for clustering an image database based on level one similarity using a new technique called the "enumeration of gradient states". This technique is based on the direction of the gradient (converting the gradient into pixel moments and computing a value known as the "gradient spin excess" for determining the complexity level of an image). This complexity level was used for clustering images into similarity groupings. From this similarity grouping or clustering, level one similarity retrieval was improved by searching each cluster for the proper membership in a cluster rather than searching the whole database. This article expands the previous study with a theoretical discussion showing that complexity based clustering using gradient spin excess is directly related to the degree of randomness (entropy) of pixel moments. In addition, we propose an improved gradient states methodology by calculating the pixel moments of an image at various sub-block sizes and clustering the image database based on hierarchical clustering using level one similarity. Finally it is shown theoretically as well as experimentally that the speed of similarity retrieval is of complexity O(n), a definite improvement over the traditional color histogram (L1-norm) similarity retrieval method.

Automated iris recognition is a promising method for noninvasive verification of identity. Although it is noninvasive, the procedure requires considerable cooperation from the user. In typical acquisition systems, the subject must carefully position the head laterally to make sure that the captured iris falls within the field-of-view of the digital image acquisition system. Furthermore, the need for sufficient energy at the plane of the detector calls for a relatively fast optical system which results in a narrow depth-of-field. This latter issue requires the user to move the head back and forth until the iris is in good focus. In this paper, we address the depth-of-field problem by studying the effectiveness of specially designed aspheres that extend the depth-of-field of the image capture system. In this initial study, we concentrate on the cubic phase mask originally proposed by Dowski and Cathey. Laboratory experiments are used to produce representative captured irises with and without cubic asphere masks modifying the imaging system. The iris images are then presented to a well-known iris recognition algorithm proposed by Daugman. In some cases we present unrestored imagery and in other cases we attempt to restore the moderate blur introduced by the asphere. Our initial results show that the use of such aspheres does indeed relax the depth-of-field requirements even without restoration of the blurred images. Furthermore, we find that restorations that produce visually pleasing iris images often actually degrade the performance of the algorithm. Different restoration parameters are examined to determine their usefulness in relation to the recognition algorithm.

Wireless sensor networks allow detailed sensing of otherwise unknown and inaccessible environments. While it would be beneficial to include cameras in a wireless sensor network because images are so rich in information, the power cost of transmitting an image across the wireless network can dramatically shorten the lifespan of the sensors. This paper investigates various compression techniques and what the cost of these algorithms would be to the lifespan of the sensor nodes. We further describe a new paradigm for cameras and wireless networks. Rather than focusing on transmitting images across the network, we show how an image can be processed locally for key features using simple detectors. Contrasted with traditional event detection systems that trigger a an image capture, this enables a new class of sensors which uses a low power imaging sensor to detect a variety of visual cues.

In order to provide good QoS for video streaming in error-prone environments, effective error control methods are essential. Current error control methods can be classified into two categories: 1) Transport layer approaches such as FEC and retransmission; 2) Application layer approaches such as error resilience coding and error concealment. By far, most existing research is aimed towards optimizing one of the above approaches to reduce the impact of transmission errors. However, there is usually more than one error control method in a real video streaming system. In this case, how to optimize the system performance becomes more complicated, and is not standardized. This paper presents the research effort to joint-optimize the effects of two error control methods, retransmission and error concealment, in wavelet-based video streaming system. The major difficulty of the joint-optimization is that the two methods are mutually dependent; the system cannot be optimized by improving each error control method independently. To tackle this problem, a new content index, namely "reconstruction distortion", is defined to quantify both the packet content and its importance in error concealment. Based on the defined content index, a content-based retransmission approach is developed to select the best packet-sending scheme to maximize the quality of the received video under the given error concealment method. Experiments results demonstrate the effectiveness of the proposed method.

In this paper, we present an efficient compression technique that is suitable for image/video communications over wireless (mobile) channel. Our technique uses basic directional differences operators to estimate corresponding detail subband images/videos from their approximation subband images/videos. We empirically found that the detail subband images/videos can be well approximated by the estimate subband images/videos. In this work, image and video are first decomposed using integer wavelet packet transformation. Having established that detail subband images/video can be estimated from the approximation subbands, the information needed to send over the wireless channel is only the most important subband images/video, where we selected them via best basis selection algorithm. Next, after best basis selection, the selection subband components are encoded using either SPIHT (JPEG) for image or 3-D SPIHT for video and then the encoded data are sent over the wireless channel. The advantages of our algorithms are two folds. First, most of the computation used in our technique is performed in integer for the purpose of coding speed improvement. Second, the computation of our algorithm (either SPIHT (JPEG) or 3-D SPIHT) is reduced from its original computation by an order of magnitude. The reason is that in our algorithm either SPIHT (JPEG) or 3-D SPIHT is performed only on the set of important components (two or a few subband image/videos) instead of the whole image/video. Finally, we show that our proposed algorithm using SPIHT (3-D SPIHT) are better that pure JPEG (MPEG-2) both in terms of human visual image and computation complexity.

Detection of intrusions for early threat assessment requires the capability of distinguishing whether the intrusion is a human, an animal, or other objects. Most low-cost security systems use simple electronic motion detection sensors to monitor motion or the location of objects within the perimeter. Although cost effective, these systems suffer from high rates of false alarm, especially when monitoring open environments. Any moving objects including animals can falsely trigger the security system. Other security systems that utilize video equipment require human interpretation of the scene in order to make real-time threat assessment. Shape-based human detection technique has been developed for accurate early threat assessments for open and remote environment. Potential threats are isolated from the static background scene using differential motion analysis and contours of the intruding objects are extracted for shape analysis. Contour points are simplified by removing redundant points connecting short and straight line segments and preserving only those with shape significance. Contours are represented in tangent space for comparison with shapes stored in database. Power cepstrum technique has been developed to search for the best matched contour in database and to distinguish a human from other objects from different viewing angles and distances.

This paper introduces an approach for synthesizing natural textures, with emphasis on quasi-periodic and structural textures. The process consists of two stages. In the first stage, the basic textural elements (texels) and the basic textural structure are determined. This is achieved by identifying two fundamental frequencies in the texture, for two different orientations. The basic structure is a non-regular mesh that defines the place holders for texels. We call such place holders e-texels (empty texels). In the second stage, a new textural structure is designed from the original one, and its e-texels are filled in by texels obtained from the original patch. Same texture texels are expected to possess a high degree of similarity, thus the new structure could be filled in at random. However, a transition probability approach is used in order to retain local textural characteristics. More specifically, assuming that texel A is the last texel placed in the new structure, the e-texel closest to A is found. The e-texel is replaced by texel B from the old structure if the relative position between A and the e-texel is similar to the relative position between A and B in the old structure. This technique is an extension of a general texture synthesis technique previously developed by the author. The proposed technique is suited for structural textures since blockage effects are eliminated by allowing irregular shape texels to be merged, contrary to the previous general technique where the blocks merged are squares. Results show that the proposed method is successful in synthesizing structural textures.

Pattern matching is one of the well-known pattern recognition techniques. When using points as matching features, a pattern matching problem becomes a point pattern matching problem. This paper proposes a novel point pattern matching algorithm that searches transformation space by transformation sampling. The algorithm defines a constraint set (a polygonal region in transformation space) for each possible pairing of a template point and a target point. Under constrained polynomial transformations that have no more than two parameters on each coordinate, the constraint sets and the transformation space can be represented as Cartesian products of 2D polygonal regions. The algorithm then rasterizes the transformation space into a discrete canvas and calculates the optimal matching at each sampled transformation efficiently by scan-converting polygons. Preliminary experiments on randomly generated point patterns show that the algorithm is effective and efficient. In addition, the running time of the algorithm is stable with respect to missing points.

This study investigates the use of thematic class correspondence in the fusion of hyperspectral data with higher spatial resolution synthetic aperture radar (SAR) data. A thematic map derived from the SAR imagery is used to introduce spatial information into the hyperspectral imagery, a spatial-spectral fusion. Because the underlying physical processes measured by the imaging systems substantially differ, classes derived from one may have partial or no relationship to classes from the other. In our approach, SAR-derived class contributions to a mixed hyperspectral pixel are weighted in the fusion process based on their correspondence with spectral classes. Unconstrained and weighted least squares solutions for the resulting linear system are described. A comparison of fusion results is presented with and without use of thematic content.

In this article we present a generic, flexible, scalable and robust approach for an intelligent real-time forensic visual system. The proposed implementation could be rapidly deployable and integrates minimum logistic support as it embeds low complexity devices (PCs and cameras) that communicate through wireless network. The goal of these advanced tools is to provide intelligent video storage of potential video evidences for fast intervention during deployment around a hazardous sector after a terrorism attack, a disaster, an air crash or before attempt of it. Advanced video analysis tools, such as segmentation and tracking are provided to support intelligent storage and annotation.

Because a signal can often be easily corrupted during its transmission, registration, or storage, de-noising is an important field in the areas of communications systems and of signal and image processing, especially where defense and security applications are of concern. Techniques employing transform-based methods such as the Fourier transform, the cosine transform, and wavelets have already been applied successfully to this field when dealing with an image corrupted by noise having a Gaussian or uniform distribution. However, images where impulse or salt and pepper noise are introduced are typically treated using median or switched-median algorithms because the sudden discontinuities of impulse noise often present problems for conventional transform-based noise reduction approaches. Additionally, binary images cannot easily be de-noised by fast orthogonal transforms or wavelets. A novel noise detection and reduction scheme using a fast logical (binary) transform-based Boolean minimization algorithm is presented. The presented approach is capable of de-noising both binary and multivalued images corrupted by impulse noise. A comparison with well-known methods is offered. Particularly, the algorithm reliably detects noise more effectively than existing switched-median methods, and de-noising results comparable to or better than those attainable with median filtering are possible. The technique performs especially well when operating on images of high complexity. The new technique does not require the use of a multiplication nor a sorting operation. In addition, we show that the presented de-noising procedure could be easily performed on an already compressed file or during the compression step. Furthermore, the simplicity of the transform makes a gate-level hardware realization practical for use with distributed sensors and inexpensive or high-speed imaging systems.

Noise is the primary visibility limit in the process of non-linear image enhancement, and is no longer a statistically stable additive noise in the post-enhancement image. Therefore novel approaches are needed to both assess and reduce spatially variable noise at this stage in overall image processing. Here we will examine the use of edge pattern analysis both for automatic assessment of spatially variable noise and as a foundation for new noise reduction methods.

Measures of image quality based on sensitivity of edge placement have previously been presented for use in imaging system analysis. Successful applications of these measures have included sharpening filter design, interpolator design, system focal length selection, compression bitrate selection, and phase diversity optical control analysis. These are applications for which the General Image Quality Equation (GIQE) is not recommended. The GIQE is intended only for assessment of optimized system designs, and is not robust in system optimization applications. The alternative to system optimization by engineering metric methods is optimization by human assessments of simulated system design alternatives, a process which is slow and expensive as well as presenting considerable practical difficulties in validation and reproducibility. A practical approach to system design combines metric evaluation of day-to-day problems for which quick answers are needed, with simulation and human evaluation of overall system performance and larger system tradeoffs. In this combined approach one use of metric analysis is to provide reasonable design alternatives to include in the trade space being explored by analyst assessments. Analysis of simple parametric studies taken from Optical Engineering is presented here in terms of the metrics, with comparison of results achieved with human analysis against results predicted from engineering metrics.

Understanding signal and noise quantities in any practical computational imaging system is critical. Knowledge of the imaging environment, optical parameters, and detector sensitivity determine the signal quantities but often noise quantities are assumed to be independent of the signal and either uniform or Gaussian additive. These simplistic noise models do not accurately model actual detectors. Accurate noise models are needed in order to design optimal systems. We describe a noise model for a modern APS CMOS detector and a number of noise sources that we will be measuring. A method for characterizing the noise sources given a set of dark images and a set of flat field images is outlined. The noise characterization data is then used to simulate dark images and flat field images. The simulated data is a very good match to the real data thus validating the model and characterization procedure.

Classical segmentation algorithms subdivide an image into its constituent components based upon some metric that defines commonality between pixels. Often, these metrics incorporate some measure of "activity" in the scene, e.g. the amount of detail that is in a region. The Multiscale Retinex with Color Restoration (MSRCR) is a general purpose, non-linear image enhancement algorithm that significantly affects the brightness, contrast and sharpness within an image. In this paper, we will analyze the impact the (MSRCR) has on segmentation results and performance.

Visual processing is a psychophysical process. In this paper, it is shown that texture processing by the human visual systems shows nonlinear dynamic patterns. This is done by developing recursion equations that describe the dynamics in the perception and maps input texture patterns to output observables. The recursion variables are complex where the real and imagery parts signify the signal and noise parts of the perception process, respectively. The output of the dynamics is used to perform texture discrimination between pairs of textures. These results are compared with results from experiments with human subjects. A good correlation is obtained between the results using texture dynamics and that of experimental data. This shows that texture perceptual dynamics is an important ingredient in developing computer algorithms that preserve perceptually similarity with human visual system processing of texture. Based on this an algorithm is developed to perform classification of textures. The algorithm is applied to classify natural and synthetic textures. This algorithm shows an improvement in the classification accuracy over traditional methods.

Shot noise is fundamental to photon detection. In image sensors there is an opportunity for incorporating lateral processing for reduction of both shot noise and thermal noise. Based on the Bayesian argument, we derive a noise-smoothing model that suppresses noise while preserving image discontinuities due to scene structure. Further, we show a possible focal plane solver of this model using a compact electronic network. Simulated experimental results are presented and similarities with human vision are discussed.

The problem of using light stripe projection (LSP) for 3D surface reconstruction is addressed in this paper. By using an adaptive filter, we show that we can recover 3D points that normally would go undetected due to light reflection and shape. Further, we show that the filter improves the accuracy of the 3D point coordinates. The filter is based on polynomial and parabola fitting. It generates a bounded polynomial or a smooth parabolic function based on a peak curve from a stripe sample and it adapts the smooth function so that other stripe sample pixels fit a new stripe sample. We then show how the filter is designed to correct the reflection affective pixels of a stripe sample and how it can improve the edge position extracted by any common edge detection method. The effectiveness of the missing 3D point recovery and the 3D point position accuracy improvement is demonstrated by the presentation of experimental results obtained using the methods described in the paper. A test demonstrating the differences between 3D point models generated with and without the adaptive filter is also presented.

Water level model is an effective method in density-based classification. To improve the result, we use biased sampling, local similarity and popularity as preprocessing, and then apply the water-level model for classification. Biased sampling is to get some information about the global structure. Similarity and local density are mainly used to understand the local structure. In biased sampling, images are divided into many l*l patches and a sample pixel is selected from each patch. Similarity at a point p, denoted by sim(p), measures the change of gray level between point p an its neighborhood N(p). Besides using biased sampling to combine spectral and spatial information, we use similarity and local popularity in selecting sample points. A sample point is chosen based on the minimum value of sin(p) + [1-P(p)] after normalization. The selected pixel is a better representative, especially near the border of an object. Kernel estimators are employed to obtain smooth density approximation. The water-level model is relatively easy and effective when the density function is smoothed. To make it more effective in other cases, one has to deal with small spikes and bumps. To get rid of the small spikes, we establish a threshold ê[f(P1) - f(P 2)*(P1-P 2) ê > c*l*l , where c is a constant, P1 is a local maximum point to be tested and P2 is the nearest local minimum form P1. The condition is only related to the size of the patches l*l. After using the average filter, we choose l to be the square root of the fifth peak if it is between 5 and 20, otherwise set l = 10. Preliminary experiments have been conducted using proposed methods with different values of the constant c in the threshold condition. Experimental results are provided.

Compression of noisy imagery usually consists of two stages, prefiltering followed by encoding. In this paper we present a technique based on on vector quantization, which combines noise reduction and compression into one step. The idea is to generate a codebook, consisting only of clean image data, which is then used for quantization of the noisy imagery. Simulations performed shows that this approach can efficiently handle images corrupted by noise, and compared to MPEG-4 encoding, this technique, in spite of its simplicity, is the better choice when dealing with high levels of noise.

We reduced speckle noise in SAR imagery by retaining only those wavelet coefficients with significant third-order correlation coefficients. These coefficients were generated from the cross-correlation functions of the image and wavelet basis functions. Using this approach, we compared the results between directly applying our denoising method, and first preprocessing by taking the logarithm of an image. In our approach, we examined wavelet coefficients in an environment where the contribution from the second-order moment of the noise had been reduced.

A robust wireless video transmission scheme using MPEG video Markup Language (MPML) with adaptive error correction codes (ECCs) is proposed in this work. MPML can provide selective description for various MPEG video components to meet the special application requirements. We present an efficient MPML compression algorithm to reduce the size of MPML description and apply the adaptive ECCs to protect the MPML-coded description bitstream unequally against channel noises. It is demonstrated by experimental results that the proposed MPML protection can achieve good error resilient performance comparable with the well-known error resilient techniques.

Burst packet loss imposes significant quality degradation for streaming applications. Interleaving, which helps reduce the probability of losing adjacent packets, is considered an effective method to mitigate burst errors. Most current research on wavelet image/video streaming is focused on how to maximize the interleaving effect in the spatial or spatial-frequency domain. However, in order to achieve the best video quality, optimizing temporal interleaving is very important, especially when error concealment is present in the streaming system because an inappropriate interleaving method may have an adverse effect on error concealment. Optimization of temporal interleaving on wavelet-compressed image/video streaming has not been previously studied. In this paper a novel optimal packet interleaving method is proposed for streaming applications on burst-loss channels. The objective is to achieve the best video quality at the receiver given an error-concealment algorithm and the channel traffic conditions. The proposed method consists of two steps: 1) spatial interleaving is conducted during packetization to disperse damage resulting from packet loss; 2) temporal interleaving is applied during transmission maximize the effect of error concealment at the receiver. In addition, a new concept that addresses the needs of error concealment, namely "temporal neighbor packet distance" is defined in order to facilitate the optimization. A low computational complexity algorithm is developed to satisfy the requirement of real-time transmission. Experimental results show that our proposed method can consistently improve the effects of error concealment.

Adaptive smoothing (AS) has been previously proposed as a method to smooth uniform regions of an image, retain contrast edges, and enhance edge boundaries. The method is an implementation of the anisotropic diffusion process which results in a gray scale image. This paper discusses modifications to the AS method for application to multi-band data which results in a color segmented image. The process was used to visually enhance the three most distinct abundance fraction images produced by the Lagrange constraint neural network learning-based unmixing of Landsat 7 Enhanced Thematic Mapper Plus multispectral sensor data. A mutual information-based method was applied to select the three most distinct fraction images for subsequent visualization as a red, green, and blue composite. A reported image restoration technique (partial restoration) was applied to the multispectral data to reduce unmixing error, although evaluation of the performance of this technique was beyond the scope of this paper. The modified smoothing process resulted in a color segmented image with homogeneous regions separated by sharpened, coregistered multiband edges. There was improved class separation with the segmented image, which has importance to subsequent operations involving data classification.

In this paper, a new multi-view stereoscopic 3D image communication system for web-based real-time teleconferencing application is proposed by using the IEEE 1394 digital cameras, Intel Xeon server computer system and Microsoft DirectShow programming library and its performance is analyzed in terms of image-grabbing frame rate and number of views. In the proposed system, two -view images are initially captured by using the IEEE 1394 stereo camera. And then, this captured two-view data is processed and compressed by extracting the disparity data from them in the Intel Xeon server computer and transmitted to the client system through a communication network, in which the received 2-view data is reconstructed and basing on this data 16-view images are generated through the intermediate views reconstruction method for more natural 3D display in the client system. The program for controlling the overall system is developed based on the Microsoft DirectShow programming library. Some experimental results shows that the proposed system can display 16-view 3D images with a gray of 8bits and a frame rate of 15fps.

The Retinex is a general-purpose image enhancement algorithm that is used to produce good visual representations of scenes. It performs a non-linear spatial/spectral transform that synthesizes strong local contrast enhancement and color constancy. A real-time, video frame rate implementation of the Retinex is required to meet the needs of various potential users. Retinex processing contains a relatively large number of complex computations, thus to achieve real-time performance using current technologies requires specialized hardware and software. In this paper we discuss the design and development of a digital signal processor (DSP) implementation of the Retinex. The target processor is a Texas Instruments TMS320C6711 floating point DSP. NTSC video is captured using a dedicated frame-grabber card, Retinex processed, and displayed on a standard monitor. We discuss the optimizations used to achieve real-time performance of the Retinex and also describe our future plans on using alternative architectures.

Image noise cancellation is necessary to remove noise generated in communication systems or remote video conferencing systems. Processing speed has become a challenge as a consequence of the increasing image resolution, especially in visual information processing. This paper presents the design and implementation of a real-time image noise canceller. Two-dimensional least mean square (TDLMS) algorithm is employed as the adaptive filter for noise cancellation. This algorithm is modified and designed with two concurrent phases: filter coefficient adjustment phase and image noise cancellation phase, with each phase mapping into a pipeline structure, therefore achieving real-time performance. The image noise canceller is implemented using hardware description language VHDL and is prototyped on Field Programmable Gate Array (FPGA) for system reconfiguration. A data buffer is developed using SelectRAM (BRAM) embedded in a Virtex FPGA to overcome the bandwidth limitation between external memory and the noise cancellation processor. The FPGA embedded multipliers are also employed to improve the processing speed. Tested using standard images, this real-time image noise canceller could process up to 1528 frames of 256 by 256 pixel images per second and could reach up to 10.4dB signal to noise ration improvement.

Visualization techniques for simulations are often limited to statistical reports, graphs, and charts, but simulations can be enhanced through the use of animation. A spatio-temporal animation allows a viewer to observe a simulation operate, rather than deduce it from numerical output. The Route Viewer, developed by Argonne National Laboratory, is a two-dimensional animation model that animates the objects and events produced by a discrete event simulation. It operates in a playback mode, whereby a simulated scenario is animated after the simulation has completed. The Route Viewer is used to verify the simulation's processes and data, but it also benefits the simulation as an analytical tool by facilitating spatial and temporal analysis. By visualizing the events of a simulated scenario in two-dimensional space, it is possible to determine whether the scenario, or simulation model, is reasonable. Further, the Route Viewer provides an awareness of what happens in a scenario, when it happens, and the completeness and efficiency of the scenario and its processes. For Army deployments, it highlights utilization of resources and where bottlenecks are occurring. This paper discusses how the Route Viewer facilitates the analysis of military deployment simulation model results.

The purpose of this paper is to describe a physics based fire model in DIRSIG. The main objective is to utilize research on radiative emissions from fire to create a 3D rendering of a scene to generate a synthetic multispectral or hyperspectral image of wildfire. These synthetic images can be used to evaluate detection algorithms and sensor platforms. To produce realistic flame structures and realistic spectral emission across the visible and infrared spectrum, we first need to produce 3D time-dependent data describing the fire evolution and its interaction with the environment. Here we utilize an existing coupled atmosphere-fire model to represent the finescale dynamics of convective processes in a wildland fire. Then the grid-based output from the fire propagation model can be used in DIRSIG along with the spectral emission representative of a wildland fire to run the ray-tracing model to create the synthetic scene. The technical approach is based on a solid understanding of user requirements for format and distribution of the information provided by a high spatial resolution remote sensing system.