Parallel Processing Of Target Detection In SAR Image

Parallel processing is an interdisciplinary of science and it provides important support for large-scale problems in science.Synthetic Aperture Radar could work in all-time and all-weather situations, so it gets wide applications in science field. It is necessary to interpret SAR images at first time. But limited to the characters of large scene and massive data, serial processing based on single computer is difficult to satisfy the actual requirement. Therefore, to find a way that can interpret SAR images fast has practical significance.In this dissertation, three algorithms of load balancing are designed and implemented in this dissertation. The performance of the algorithms is detailed and deeply analysed in the different general parallel processing system with different target detecting algorithms.Firstly, this dissertation briefly introduces the development history of parallel processing and target detection of SAR image. Secondly, the basic theory of parallel processing and two typical kinds of target detection of SAR image is introduced. Then, on the basis of the way that convert a general problem of serial processing to parallel processing, the feasibility of target detection algorithms in SAR image based on parallel processing is discussed. After that, the way of data segmentation is designed combined with the characters of target detection in SAR image.According to the different parallel processing systems, two general algorithms of load balancing suitable for target detection in SAR image are designed. They are static load balancing and dynamic load balancing. Base on the character of parallel processing and target detecting, a new algorithm named best static load balancing on the basis of test- feedback control is proposed. Then, CFAR detector and EF detector are used to test the three algorithms of load balancing in different parallel processing systems, and the performance is explored and analyzed deeply.At last, the whole research work is summarized, and personal opinions about this study field are given.