自适应分割的雷达回波数据混合编码算法  

ADAPTIVE SEGMENTATION OF RADAR RETURN AND HYBRID CODING SCHEME

作　　者：薛建儒[1] 郑南宁[1] 权炜[1] 

机构地区：[1]西安交通大学人工智能与机器人研究所,西安710049

出　　处：《电子与信息学报》2003年第8期1059-1065,共7页Journal of Electronics & Information Technology

基　　金：国家自然科学基金(60024301)

摘　　要：该文提出了一种基于自适应分割模型的雷达信号混合编码算法。首先用自适应高斯混合体模型将雷达回波数据分为两类;一类是由运动目标组成的前景图;一类是视野中静止(如礁石等)或具有重复性运动目标(如海浪等)的背景图。然后对前景图和背景图分别用整数小波去相关,对变换后的前景图和背景图分别采用改进后的SPIHT无损压缩和有损压缩编码。实验表明,此编码算法可有效降低数据传输率,在相同传输码率下,相对于不分割直接压缩算法EZW和SPIHT,重构信号的PSNR增加0.3-0.5dB,主观质量也优于同类算法。In this paper, an adaptive segmentating, hybrid coding scheme of radar return is presented. Firstly, radar return is divided into two classes by adaptive Gaussian mixture model: foreground and background, the former is composed of moving targets, while the later is composed of still objects and repeat motion. Then Integer-to-integer wavelet transform is used for de-correlating the background and foreground, and an improved SPIHT codec is also used in coding wavelet coefficients of foreground in lossless way and wavelet coefficients of background in loss way. Experimental results show the scheme in this paper is better than codec based on SPIHT directly, and the PSNR of the reconstructed image is also increased about 0.3-0.5dB in average under the same bit rate compared with that of the latter one.

关 键 词：自适应高斯混合体 信号压缩 数据分割 整数小波变换 雷达 回波数据 

分 类 号：TN957.51[电子电信—信号与信息处理]