MODIS多光谱图像压缩研究  

Compression for MODIS Multispectral Images

作　　者：李元祥[1] 邓黎[2] 敬忠良[1] 

机构地区：[1]上海交通大学空天科学技术研究院,上海200240 [2]解放军理工大学气象学院,南京211101

出　　处：《电子与信息学报》2008年第10期2352-2355,共4页Journal of Electronics & Information Technology

基　　金：国家自然科学基金(40575010)资助课题

摘　　要：为有效存储MODIS多光谱图像数据,该文提出一种基于谱间预测和整数小波变换的多光谱图像压缩算法。首先通过构造谱间最优预测器去除谱间冗余,再利用整数小波变换和SPIHT算法对预测误差图像去除空间冗余,最后进行自适应算术编码。该方法可实现MODIS多光谱图像的无损、近无损和有损压缩,取得了满意的实验结果;在不同小波基条件下与3D-SPIHT算法比较,表明了该方法的有效性。In order to store MODIS multispectral images effectively, a new hybrid compression algorithm based on optimal inter-band prediction and integer wavelet transform is proposed. First, an optimal inter-band predictor is designed for de-correlating the spectral redundancy, and the error images are created for later use. Next, spatial redundancy is removed by using an efficient integer wavelet coding method and SPIHT algorithm. Then, the adaptive arithmetic encoder is employed for binary code stream. The proposed method can implement the lossless compression, near-lossless compression and lossy compression with high compression ratio. The experimental results show that this method is more efficient than the traditional 3D-SPIHT algorithm and other wavelet transform methods with different wavelet bases.

关 键 词：MODIS多光谱图像 SPIHT编码 最优预测 整数小波变换 

分 类 号：TP751.1[自动化与计算机技术—检测技术与自动化装置]