A synthetic aperture radar(SAR) obtainsazimuth resolutionby combiningdata from a number of points along a specified path. Uncompensatedantenna motion thatdeviates significantlyfromthe desired path produces spatially-varianterrorsin the outputimage. The algorithm presented in this paper correctsmany of these motion-relatederrors. In this respecLit is similarto time-domainconvolution,but it is more computationallyefficient. The algorithmuses overlappedsubaperturesin a three-stepimage-formationprocess: coarse-resolutionazimuth processing, fine-resolutionrange processing, and fine-resolutionazimuth processing. Rangemigrationis corrected afterthe first stage, based on coarse azimuth position. Prior to the final azimuth-compression step, data coordinatesare determined to fine resolution in range and coarseresolutionin azimuth. This coordinateinformationis combined with measuredmotion data to generate a phase correction that removes spatially-variant errors. The algorithm is well-suited for real-time applications, particularly where large flight-path deviationsmust be tolerated. 2. INTRODUCTION SAR image formation can be viewed as the process of compressing a return signal by correlating it with a reference pointtarget response. In principal, the most general SAR algorithm, time-domain convolution, computes an exact reference function and correlation for every image pixel. Other methods make"approximationsthat trade performance for reduced complexity. These various algorithms differ in how the?,'modelthe reference function and how they implement the correlation. Much of the difference in reference functions is due to operating conditions (radar frequency, aircraft vs. satellite platforms) and image requirements(size, resolution). Correlator implementation is affected by a number of factors includingcomplexity of the reference function,processing-timerequirements,andavailability of radar andsignal-processing hardware. No single image-formationtechniqueis appropriatefor all applications. The overlapped-subaperture (OSA) algorithmdiscussed in this paper w_ designed for real-time, fine-resolutionoperation in an environment where the radar platform experiences significant off-track motion. Because of the real-time aspect, the algorithm is constructedlargely with FFTand vector-multiplicationoperations. This is important since availableDSP chips can compute FFTs quickly and manipulatedata as vectors. Some motion-compensation steps are carried out before the returnsignal is digitized. This motioncompensation is implementedby changing the radar frequency, phase, PRF, and A/D sample rate on a pulse-to-pulse basis. The radarcomputes these parameters as functions of motion data measuredby an inertial navigation system. To provide fine resolution in range, the range reference function needs to vary with azimuth position as well as range. The initial stage of OSA processing filters the input data into Doppler-frequency bands corresponding to coarse azimuth position. Filtering at this point increases computational efficiency, becausei...
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.