基于视觉系统模型的SAR图像压缩方法  

Algorithm of SAR Image Compression Based on Human Vision System

作　　者：沈冰[1] 张海波[2] 郎方年[3] 朱兴华[1] 杨定宇[1] 

机构地区：[1]成都信息工程学院光电技术学院,成都610225 [2]成都信息工程学院控制工程学院,成都610225 [3]成都大学计算机学院,成都610054

出　　处：《半导体光电》2010年第4期657-660,共4页Semiconductor Optoelectronics

基　　金：国家自然科学基金资助项目(60873092)

摘　　要：提出一种基于视觉模型的合成孔径雷达(SAR)图像压缩方法,对SAR图像分解后不同频带的小波系数结合其能量分布和人眼对不同频率段的敏感度采用可变步长进行量化与压缩编码,应用嵌入式小波图像压缩算法,以峰值信噪比和均方误差等作为评价指标,对比进行了实验。结果表明,在保证压缩比不变的前提下,采用本算法获得的恢复图像的主观质量有所提高,较好地保留了图像的纹理细节,提高了图像的峰值信噪比。SAS image compression based on the model of human vision system is presented in this paper.The wavelet coefficients of the decomposed SAR images were variably quantized and coded based on the energy distribution of wavelet＇s different frequency bands and corresponding human vision sensitivity.Using the embedded wavelet image decompressed algorithm,and the evaluation parameters of PSNR and MER,experiments are carried out to make comparisions.The results show that,under the condition of the same ratio of image decompression,the proposed approach can obtain better visual quality and texture details of the images and improve PSNR of SAR.

关 键 词：SAR图像压缩 视觉系统模型 嵌入式零树小波 多级树集合划分编码 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]