四叉树分块的高光谱图像分布式无损编码  

Lossless distributed source coding of hyperspectral images based on quadtree segmentation

作　　者：王相海[1,2] 张智迪 宋传鸣[1] 

机构地区：[1]辽宁师范大学计算机与信息技术学院,大连116029 [2]辽宁师范大学自然地理与空间信息科学辽宁重点实验室,大连116029

出　　处：《中国图象图形学报》2015年第8期1102-1109,共8页Journal of Image and Graphics

基　　金：国家自然科学基金项目(41271422;61402214);高等学校博士学科点专项科研基金项目(20132136110002);辽宁省博士科研启动基金项目(20121076);辽宁省教育厅科学研究一般项目(L2013405;L2013405)

摘　　要：目的近年来随着光谱成像技术的快速发展,使得高光谱遥感图像能够提供更加丰富的地物信息,然而其所具有的极大数据量给图像的存储、传输和实用带来较大的困难。因此,如何对高光谱遥感图像进行有效编码成为研究热点。方法基于陪集码的分布式信源编码因其具有良好的压缩性能和较低的编码复杂度而受到重视,在此基础上提出一种基于自适应四叉树分块的高光谱图像分布式无损编码方案。设每个高光谱帧组的第1帧为关键帧,其他帧为普通帧,首先对关键帧进行自适应四叉树分块,然后对每一块与普通帧相应的块进行最佳线性预测,进而根据预测残差确定所需传输的陪集码索引位数以及每个像素的k个最低比特位。结果本文通过自适应四叉树分块,增强了所形成陪集码的自适应性。所提出的编码方案很好地实现了压缩效果和编码效率的折中。结论提出的方案能够较好地适应低复杂度环境下对高光谱图像无损编码的需求。Objective With the rapid development of spectral imaging technology in recent years, hyperspectral remote sensing images can provide abundant data on surface features. However, the sizable data of hyperspectral images make their storage, transmittal, and application quite difficult. As a result, how to validly code hyperspectral images has become a hot issue. Method The distributed source coding based on coset codes has received much attention because of its good compression performance and low coding complexity. In this study, we present a scheme for lossless distributed source coding of hyperspectral images based on adaptive quadtree segmentation. Assuming that the first frame in every group of spectrum frames is the key frame, the other frames are Wyner-Ziv frames. First, we perform adaptive quadtree segmentation on the key frame and then optimum linear prediction on each block of each Wyner-Ziv frame. Afterwards, the index of the eoset codes to be transferred and the k least significant bits of every pixel in this block are ascertained using prediction residuals. Result In this study, adaptive quadtree segmentation scheme has been proved to strengthen the adaptability of the formed coset codes. The proposed scheme can achieve a good compromise between coding efficiency and calculating complexity. Conclusion This scheme is better able to meet the lossless coding requirement for hyperspectral images under low-complexity environment.

关 键 词：高光谱遥感图像 分布式信源编码 四叉树分块 陪集码 分块信息记载 预测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]