基于DCT的分块自适应量化算法及其用于SAR原始数据压缩  

A Compression Algorithm for SAR Raw Data Based on the Combination of Discrete Cosine Transform and Block-Adaptive Quantization

作　　者：潘志刚[1] 王岩飞[1] 

机构地区：[1]中国科学院电子学研究所

出　　处：《电子与信息学报》2007年第8期1784-1788,共5页Journal of Electronics & Information Technology

摘　　要：该文提出了一种基于离散余弦变换(DCT)和分块自适应量化相结合的SAR原始数据压缩算法。利用SAR原始数据满足局部平稳高斯随机过程的特点,通过将DCT系数进行重排,并对重排后的系数矩阵进行有效的量化比特分配和分块自适应量化,从而大幅度提高了量化增益。通过对真实SAR原始数据的压缩实验结果表明:该文算法与BAQ算法相比,以相对较低的运算复杂度增加,使图像域的压缩性能指标有了明显提高。In this paper, an algorithm for compressing synthetic aperture radar raw data is proposed. This algorithm is based on the combination of discrete cosine transform and block-adaptive quantization. The known results that a block normalized SAR raw signal is a Gaussian stationary process are exploited in order to re-array the DCT coefficients. Coupled with a proper bit allocation strategy and block-adaptive quantization on re-arrayed coefficients matrix, this algorithm exhibits notably an interesting performance/complexity trade-off with respect to conventional methods such as BAQ. Simulation results on real-world SAR rave data also show that the proposed algorithm outperforms methods based on wavelets as to SNR and PSNR.

关 键 词：SAR原始数据压缩 离散余弦变换 比特分配 分块自适应量化 

分 类 号：TN957.51[电子电信—信号与信息处理]