Radar images are created by converting time delay and doppler measurements of scatterers into range and cross-range values. An imaging radar does not provide direct height measurements of scatterers. We investigate the problem of computing height information from a pair of radar images. Elevated scatterers will appear closer in range; this phenomenon is called radar image layover. This paper investigates how the height of a scatterer can be computed from the difference in its layover between two images. First, it is shown that for any image pair taken from a constant- altitude straight-line flight path, the difference in layover is zero and therefore images are spatially coherent. Next, an expression is derived for accuracy of height estimate as a function of range resolution and the angular difference between slant planes (image planes). As the angle between the slant planes increases, the accuracy of the height estimate improves, however, bright scatterers in one image tend to fade in the other image. This tradeoff between the accuracy of height estimate and limited angular persistence of radar scatterers is discussed. Finally, results are shown, based on Lincoln Laboratory high-resolution (0.3 meters) polarimetric SAR imagery.

The paper is about the design and working principles of an experimental polarimetric ISAR system for measuring two-dimensional polarimetric radar signatures of aircraft in dynamic operational environments. The system is based on targets equipped with a coherent transponder for generating a reference signal which upon reception is used in the receiver to compensate for non-linearity in the VCO-generated linear-FM signal, as well as for performing the required motion compensation with optimum accuracy. The system makes it possible to study 2-D polarimetric ISAR in an "ideal" situation, where the limitations are due to factors other than motion compensation. Preliminary results demonstrating the functionality of the system are included.

A modified spotlight synthetic aperture radar (SAR) mode is introduced which may reduce the signal processing load often required to correct for range- walk and range-ambiguity effects that can be present with the use of the conventional spotlight SAR mode. An illustrative millimeter wave (MMW) SAR example is given.

The Lincoln Laboratory millimeter-wave Synthetic Aperture Radar (SAR) imaging system is part of a DARPA-funded program that was established at Lincoln Laboratory to investigate the detection and classification of stationary targets using ultra-high resolution, fully polarimetric SAR and Real Aperture Radar (RAR) data. The system consists of an airborne radar that operates at 33.56 GHz. The raw radar data are recorded on high density digital tapes that are sent to the Radar Data Analysis Center, which is located at Lincoln Laboratory in Lexington, MA. This center processes the data to create calibrated SAR and RAR images. The Radar Data Analysis Center consists of a number of major data processing elements: an image formation processor, an archival storage and retrieval system, and a cluster of computer systems used for data analysis. In order to accomplish the goals of the DARPA program, it is essential that the radar data be very carefully calibrated. The calibration process consists of three major steps: (1) An internally-generated calibration pulse is inserted into the radar receiver at the front end. (2) Calibration targets (dihedrals and trihedrals) deployed on the ground are measured by the radar from the air. (3) Special calibration processing software uses the measurements from (1) and (2) to achieve polairmetric calibration. This paper will describe the airborne radar, the ground processing facility, and the calibration process. Recent SAR images, generated from airborne measurements, of ground clutter and selected urban areas will be presented. The images were generated using the polarimetric whitening filter (PWF), a novel processing technique developed at Lincoln Laboratory. The PWF process exploits the polarimetric measurement capability of the radar to create imagery the is nearly optical in quality.

The Synthetic Aperture Radar (SAR) is increasingly gaining acceptance as an important sensor with military and commercial applications. Characterization of the SAR sensor is paramount for interpreting the data and developing new applications. This paper reports on work done at the Naval Air Development Center (NADC) to quantify and correct the P3 SAR imagery data for various system effects. Among the major factors that affect SAR images is the antenna pattern. For the P3 SAR the antenna effect is particularly apparent at X band, where the imaged area is comparable in extent to the area illuminated by the main beam. The antenna pattern introduces a gain attenuation away from the main axis, as well as gain modulations due to pattern interactions with the aircraft structure. Additional factors that affect SAR images include range attenuation, range dependent processing gain, system non-linearities, and others. The paper presents integrated SAR radiometric calibration procedures developed at the NADC, that build on previous work done in this area by researchers at the Environmental Research Institute of Michigan (ERIM). The procedures enable relation and comparison of SAR imagery collected at different times and conditions, and provide for actual radiometric evaluation of SAR data. Tests were carried out to evaluate the proposed methods using targets of known radar cross section. The test results confirm the feasibility of the calibration process.

The MIT Lincoln Laboratory 33.6 GHz airborne synthetic aperture radar (SAR) is a high- resolution (0.3 m x 0.3 m), fully polarimetric instrumentation radar system1. Performance specifications for this sensor are stringent2, and in order to meet these specifications, sophisticated calibration procedures have been developed and are implemented during ground-based, post-mission data processing. This paper reviews the performance specifications, describes the calibration procedures, and provides a status of the calibration effort since the sensor became operational in 1988.

In this paper basic principle of range-Doppler imaging in inverse synthetic aperture radar (ISAR) is analyzed. Simulating experiments of ISAR imaging on far-distant rotating multi-point targets are finished. The data films of ISAR in rectangular and polar coordinate formats are generated by microcomputer. Two-dimensional image with high- resolution are obtained by optical processing technology. Possible motion errors are discussed in the paper. Compensating methods and experimental results for some motion errors are given. Finally, feasibility of optoelectronic hybrid real-time processing for ISAR data is investigated.

One of the methods that can be used to enhance stationary target detection performance is to combine radar data from several looks at an area that may contain targets. This paper presents a study of several multilook techniques. The data used in the study were collected using the MIT Lincoln Laboratory 33.6 GHz Synthetic Aperture Radar (SAR) in the spotlight mode; this mode maintains the radar beam on the same area as the aircraft flies by. Consecutive 0.3 m by 0.3 m resolution images were registered to a single coordinate frame, and then combined in various ways. The processing techniques studied included some methods that combine the data prior to detection (such as noncoherent averaging, which reduces speckle), and others- that combine the detections from individual images (such as techniques that require m detections in n images).

An efficient method for visual data compression is presented, combining generalized Peano Scan, wavelet decomposition and adaptive sub-band coding technique. The Peano Scan which is an application of the Peano curve to the scanning of images, is incorporated with the encoding scheme in order to cluster highly correlated pixels. Using wavelet decomposition, an adaptive sub-band coding technique is developed to encode each sub-band separately with an optimum algorithm. Discrete Cosine Transform (DCT) is applied on the low spatial frequency sub-band, and high spatial frequency sub-bands are encoded using Run Length encoding technique.

Radar images are created by converting time delay and doppler measurements of scatterers into range and cross-range values. An imaging radar does not provide direct height measurements of scatterers. We investigate the problem of computing height information from a pair of radar images. Elevated scatterers will appear closer in range; this phenomenon is called radar image layover. This paper investigates how the height of a scatterer can be computed from the difference in its layover between two images. First, it is shown that for any image pair taken from a constant- altitude straight-line flight path, the difference in layover is zero and therefore images are spatially coherent. Next, an expression is derived for accuracy of height estimate as a function of range resolution and the angular difference between slant planes (image planes). As the angle between the slant planes increases, the accuracy of the height estimate improves, however, bright scatterers in one image tend to fade in the other image. This tradeoff between the accuracy of height estimate and limited angular persistence of radar scatterers is discussed. Finally, results are shown, based on Lincoln Laboratory high-resolution (0.3 meters) polarimetric SAR imagery.

We investigate a type of artificial neural network which has been called a high order network for application to the millimeter wave (MMW) radar stationary target classification problem. The high order network like the multilayer perceptron provides a minimum mean square error (MMSE) estimate of the optimal discriminant, however, the high order network has the advantage of ease of training. This network can be trained via iterative gradient descent and also by a closed form one-pass solution. Using real beam Ka-band radar field data, we compare the classification performance of the high order network with that of a gaussian classifier for several conditions. We found that the high order network performance is improved over the gaussian classifier and further, we obtained very attractive results with the one-pass solution.

This paper consists of two parts. The first addresses the formation of 2-D high resolution ISAR images of targets such as aircraft and other vehicles without a priori knowledge of target trajectory. The same procedures provide SAR imaging without electromechanical platform stabilization. Two data- adaptive self-calibration algorithms that perform beamforming and focusing are described. The radio camera is the self-calibrating phased-array instrument used. X-band, 150 MHz bandwidth images are presented. The second part shows that SAR detection sensitivity is considerably enhanced when stereo pairs of such images are presented to the operator. Experiments are described using radar data of a commercial airplane for the target, and fields of farmland for the clutter background. The images of each stereo pair differ only in the horizontal locations of the targets relative to the clutter. Each experiment was viewed by eight observers. The probability of target detection (Pd) vs. target to clutter ratio (TCR) was measured in the first experiment, in which the stereo viewer was an old-fashioned hand-held stereoscope. The use of stereo pairs, relative to single-image viewing, increased detection sensitivity by 7 dB. In the second experiment the stereoscope was replaced by a sophisticated computer-driven electronic display and the threshold of target detectability was measured. The increase in detection sensitivity, again relative to single-image viewing, varied between 20 and 40 dB. The anticipated gain under nonlaboratory conditions is the order of 20 dB.

This paper describes several techniques for enhancing fully polarimetric, high-resolution synthetic aperture radar (SAR) imagery. Image enhancement is viewed as a pre-processing step to prepare the SAR data for sophisticated detection, discrimination, and classification algorithms. The paper considers enhancement techniques that use polarimetric information in the imagery to achieve one or both of the following goals: (1) reduction of image speckle, and (2) improvement in target-to-clutter contrast. Three enhancement techniques are presented: the polarimetric whitening filter (PWF), the polarimetric enhancement filter (PEF), and the polarimetric matched filter (PMF). These polarimetric processing techniques are applied to actual SAR data gathered by the Lincoln Laboratory MMW airborne sensor. The resulting target and clutter statistics are compared for typical data sets at both 1ft by 1ft resolution and lm by lm resolution.

The nonparametric classification scheme called learning vector quantization, LVQ, is applied to the detection of an object in clutter with SAR imagery. The LVQ structure self-organizes into a decision mapping via unsupervised training which can require a relatively large data set. A physical model can generate a training set for the LVQ prior to further training with an actual data set. The approach is especially attractive when the physical model is too complex to yield an optimal, or near optimal, decision rule. Here a two-scale electromagnetic scattering model is used derive a SAR image model that includes obscuration, shadowing and range inversion. The LVQ performance achieved is good and comparable to the optimum in cases permitting analysis.

While radar interferometry has found application in areas as diverse as the study of ocean currents and generation of digital elevation maps, few studies of its potential to detect and quantify surface change have been performed. When an imaging platform makes several passes over a given site, a number of different interferograms representing different temporal and spatial baselines can be generated. The complex backscatter from a particular target can then be compared for the various images, and the effects due to the spatial baselines and noise levels removed. What remains is a measure of the ’’decorrelation” of the radar signals with time, which is indicative of changes in the surface occuring during the period of time spanned by the images.

A high PRF rate single frequency instrumentation radar is used to generate two-dimensional images of rotating structures such as helicopter blades and hubs. The technique utilizes a tomographic backprojection reconstruction algorithm where Doppler spectra generate cross-range one-dimensional profiles of the rotating structure. A simplified computer generated scattering model is developed to show proof-of-concept and aid in developing the reconstruction algorithm.

A new “product” clutter model is proposed for fitting the cumulative distribution function of high- resolution SAR data. Results shown in this paper indicate that this new clutter model fits the measured clutter data more accurately than the traditional product clutter models.

Calibration of Synthetic Aperture Radar (SAR) data is essential it is to be used in quantitative studies of the Earth's surface. At the Jet Propulsion Laboratory, we have developed a number of approaches for calibrating SAR image data, including data from multi-polarization and multi-frequency systems. This has resulted in an archive of calibrated data at JPL from several different types of scene. In this paper, it will be shown how calibrated SAR data can be used to improve our understanding of some physical properties of the Earth's surface layer. We will show how calibration uncertainties and the presence of system noise should be handled by the SAR data user. Using examples of calibrated SAR data from the NASA/JPL DC-8 SAR, it will be demonstrated how calibrated data can be used to monitor temporal change and to improve the classification of land cover type.

Full polarimetric turntable target signature data is used to develop Inverse Synthetic Aperture Radar (ISAR) plots to analyze high resolution target scatter centers. The data was collected using in-scene calibration reflectors to correct for transmit and receive polarization distortions in amplitude and phase. Scatter centers identified to have high radar cross section (RCS) with strong persistency are then analyzed in full polarimetric scatter matrix (PSM) space to discover uniqueness between target types. Target decomposition techniques are used to analyze the target centers. A unique scheme for isolating the effects of a desired scatter phenomenology is described in this paper with results shown for a truck target.

Phase and amplitude fluctuations induced by wave propagation through foliage limit the ability of a synthetic aperture radar (SAR.) system to image a target under foliage. One-way measurements of these fluctuations were done using the NASA/JPL C-, L-, UHF band SAR during the Lincoln Laboratory July 1990 Foliage Penetration Experiment. In this experiment, single-frequency CW signal sources, “tone generators,” were placed in the open and under foliage in order to measure one-way propagation. Three tone generator sites located under trees and one tone generator site in the open were utilized. Tone generators at the open site were used as controls. At each site, six tone generators were deployed, one for each frequency and polarization utilized by the NASA/JPL SAR system. The statistical properties of the phase and amplitude fluctuations induced by the foliage are determined for each tone generator site, frequency, and polarization. The effect of these amplitude and phase fluctuations on the ability of a SAR system to image a target obscured by foliage is shown. These results are compared with the measured ground truth for each tone generator site. The ground truth measured during the experiment includes foliage densities, moisture contents, and permittivities.