Real-time Compression Of Synthetic Aperture Radar Image Data

Synthetic Aperture Radar is a high-resolution microwave imaging radar,usually installed in planes or satellites for the use of imaging the land. Theimages formed are thus characterized by huge size and enormous data. Due tothe development of modem SAR system, namely, the appearance of techniqueslike high-resolution, wide mapping swath, multi-frequency, multipolarizationand multimode, greater challenges have to be confronted in the field of datatransmission and store. High quality compression arithmetic, made feasible bythe support of the developing real-time imaging system, could well be applied toeffectively deal with the problem in data transmission and store.These years, Wavelet Transform (WT) is widely used in image dealing.The essence of WT is multiresolution analysis, which fit human's visual systemvery well, and WT has good local characters both in time domain and frequencydomain, so this paper mainly focuses on the design of real-time compressionarithmetic based on WT and the implementation of this arithmetic in hardwarecircuitThe main work and innovation are as follows:(1) Elaborate on the Wavelet theories and general compression methodsbased on Wavelet Transform. Research Embedded Zerotree Wavelet (EZW)arithmetic and Set Partitioning in Hierarcical Trees (SPIHT) arithmetic,both of which are based on the Spatial-Orientation Tree(SOT) architecture.Introducing the improved SPIHT arithmetic, which can reduce thememory used in the process of encoding and help its implementation.(2) Propose a new arithmetic called Layered Quantization Of HierarchicalTrees Based On Statistics(BS-LQOHT), aimed at main requests of SARimage real-time compression. This arithmetic can change threshold inaccordance with image data and layer quantizate the object. Compared toSPIHT arithmetic, this arithmetic can effectively reduce the memory usedin the process of encoding and greatly improve processing speed, with no loss in peak signal-to-noise ratio(PSNR). Consequently, it is of higherpractical value for big image like SAR image.(3) Implement the arithmetic. A fixed-point Digital Signal Processing chipcalled TMS320C6415 produced by TI has been used in this circuit. Itshigh speed and precision can meet the demand of real-time processing ofSAR image data compression. Furthermore, this circuit has used a highperformance connector which conforms to the CPCI specification.(4) Research two problems encountered in software developing process: 1).the boundary processing of wavelet transform; 2).calculation of thenumber of bits to output one time in improved SPIHT arithmetic; andconfirm the circuit can compress the SAR image data in time...