块自适应预测量化算法压缩SAR原始数据

分析了使用块自适应预测量化(block adaptive predictive quantization,BAPQ)算法压缩SAR原始数据,分别讨论了使用均匀量化器、Lloyd-Max量化器及阈值量化器时,BAPQ算法的性能情况。提出的使用阈值量化器量化输入信号与预测信号差值信号的算法,可以充分利用差值信号接近零值个数增多以及阈值量化器量化高斯分布信号性能更优的特点,增加压缩系统的性能。通过进一步分析BAPQ算法的复杂度,认为BAPQ算法在低编码比特率条件下是算法性能与算法复杂度之间较好的折中,能适应现代SAR系统数据压缩的需求。

Block adaptive predictive quantization for SAR raw data compression

The block adaptive predictive quantization(BAPQ) algorithm for SAR raw data compression is proposed.The performances of the BAPQ with uniform quantizer,Lloyd-Max quantizer and threshold quantizer are evaluated,respectively.The result is exploited that many prediction error signals are close to zero and few of them have large amplitude.The employment of the threshold quantizer optimized for Gaussian statistics and the BAPQ using threshold quantizer can obtain a higher performance.With the analyses of algorithm's complexity,the proposed BAPQ algorithm shows a good performance/complexity trade-off,and is particularly(well-suited) to SAR raw data compression.