大孔径静态干涉光谱仪图像压缩技术  

The Image Compression Technology of the Large Aperture Static Imaging Interference Spectrometry

作　　者：肖江[1] 周有喜[1] 吴成柯[1] 杨建峰[2] 相里斌[2] 

机构地区：[1]西安电子科技大学综合业务网国家重点实验室,西安710071 [2]中国科学院西安光学精密机械研究所,西安710068

出　　处：《光学学报》2004年第11期1494-1498,共5页Acta Optica Sinica

基　　金：十五军事通信技术预研 (4 10 0 10 30 2 )资助课题

摘　　要：从图像压缩的角度分析了大孔径静态干涉成像光谱仪成像原理 ,并提出了一种新的遥感多光谱图像压缩方案。新方案利用大孔径静态干涉成像光谱仪推扫成像特点提出了一种低存储量 ,帧间小波域匹配的图像序列压缩方法 ,提高图像质量 3～ 4dB。为了保证图像的光谱信息在 8倍压缩比下有效应用 ,系统采用了一种新的感兴趣区域 (ROI)编码技术。感兴趣区域编码时 ,由于方案中采用率失真优化斜率提升 ,而不是比特平面移位 ,从而使图像在相同的光谱分辨率下拥有更好的空间分辨率。实验结果表明 ,算法大大保护了图像的光谱特性 ,在 8倍压缩比下 ,满足了该类干涉多光谱遥感图像的质量要求。A new compression scheme of interferential multispectral images is proposed based on the analyses of large aperture static imaging spectrometry. There are two novelties in the scheme: the first is that the compression coding is an image sequence compression method with low memory, which utilizes the matching algorithm in the wavelet domain to remove the redundancy between the two successive frames; The second is that in order to protect the information of the spectrum, a new ROI (Region of interest) coding algorithm is adopted, in which the lifting procedure is carried out in code stream organization stage instead of the DWT as suggested by JPEG2000. As a result, the low coding efficiency situation due to the nature of the bitplane shift methods could be improved, which enables the high fidelity in the background (BG) and makes better qualities of spectrum curve. The experimental results have been proved to be encouraging.

关 键 词：比特平面 感兴趣区域编码 图像压缩技术 率失真优化 编码技术 压缩比 小波域 多光谱遥感图像 多光谱图像 图像序列 

分 类 号：O436[机械工程—光学工程] TN919.81[理学—光学]