全局约束自适应量化的遥感图像压缩算法  

Algorithm for satellite remote sensing image compression with adaptive quantization based on global restrictions

作　　者：季玲玲[1] 陈浩[1] 张晔[1] 

机构地区：[1]哈尔滨工业大学信息工程系,哈尔滨150001

出　　处：《哈尔滨工业大学学报》2009年第5期15-20,共6页Journal of Harbin Institute of Technology

基　　金：国家自然科学基金资助项目(60472048)

摘　　要：针对JPEG2000算法存在大量计算冗余影响星载遥感图像压缩速度的问题,提出一种全局约束自适应量化的星载遥感图像快速压缩算法.根据给定目标码率、输入图像特性和各子带特性,考虑JPEG2000内部之间的联系,从全局角度自适应设置各子带量化步长,从而减少最为耗时的EBCOT待编码数据量.实验结果表明,该算法减少了JasPer90%-95%的冗余数据,从而减少了层一编码时间和层二截断时间,显著提高了JPEG2000的编码速度,相比JasPer速度提高了1倍左右,而重构图像质量还略有提高.To deal with the redundant computation in JPEG2000 which affects the compression speed of satellite remote sensing image. An algorithm for fast compression of satellite remote sensing image with adaptive quantization based on global restrictions（AQGR） is proposed . According to the target bit rate, input image and the characteristics of subbands, the subband quantization step is set up adaptively from the overall perspective with consideration of inner relationship in the processes of JPEG2000 witch reduces redundant data which are the most time-consuming to be encoded in EBCOT. Experimental results show that compared with Jasper, AQGR reduces 90% -95% redundant data, thereby reduces the encoding time in Tilel and truncation time in Tile2, and significantly improves the coding speed of JPEG2000. Furthermore, the peak signal -tonoise ratio（PSNR） of the reconstructed image is increased slightly.

关 键 词：JPEG2000标准 遥感图像压缩 自适应量化 全局约束 

分 类 号：TN911.73[电子电信—通信与信息系统]