Study And Implementation Of Real-time ISAR Imaging Technology Based On ADSP-TS201S

Inverse synthetic aperture radar (ISAR) is a kind of high-resolution imaging radar, which can be able to obtain high-resolution two-dimensional images of the targets in a great distance. It provides important instruments for all-weather and all-time identifying of long-range targets.With the development of science and technology, and the urgent requirements of military struggle, the real-time ISAR imaging systems become an important research.This paper studies real-time ISAR imaging algorithm. According to the design process of parallel algorithm of multi-DSP, based on the hardware platform which consists of six ADSP-TS201S, this paper designs the software of real-time ISAR imaging and in order to improve real-time, three levels of parallel processing technology and four methods of real-time improvements are studied and realized.Firstly, the theories of parallel processing techniques are described. The three factors: processing units, interconnect structure, allocation of algorithms and tasks, which can affect the capability of parallel processing, are analyzed. Then, the real-time ISAR imaging algorithm is studied. Aimed at two issues of traditional range alignment algorithms, global cross-correlation method based on index weighted which can reduce the alignment error is studied and three methods,including interpolation and decimation, polynomial fitting, the way that searches through sub range units, are studied too.Based on real measured data, according to three indicators which are image entropy ,image contrast and real-time, a comprehensive comparison of various range alignment algorithms is accomplished. Four kinds of phase compensation algorithms are analyzed, and a comparison of the entire four algorithms based on the real-time and ISAR imaging quality is executed. According to the performance of all different algorithms above, the range alignment algorithm and phase compensation algorithm which are suitable for real-time ISAR imaging are selected.With the completion of requirements analysis, task partitioning, task allocation and the communication design, the software of real-time ISAR imaging is designed, based on hardware platform. Three levels of parallel processing are applied, including the parallelism within instructions, parallelism between instructions and the task-level parallelism. Four ways of real-time improvements, consisting of sliding window, matrix transpose, fast algorithm of cross-correlation and pipeline expansion are proposed and a method to implement sliding window and matrix transpose at the same time is put forward. This paper ultimately realizes real-time ISAR imaging with high performance. According to actual demand, a boot program for six DSP which mixes EPROM boot and LINK boot designed and implemented.Based on real measured data, the treatise summarizes the characteristics and applicability of each algorithm of range alignment and phase compensation. The conclusions are of great value for other radar systems to design and select the imaging algorithm. The real-time improved methods which are proposed in the treatise are useful because the core of real-time imaging software has good portability.