基于整数小波变换的MODIS多光谱图像无损压缩  

Lossless compression based on integer wavelet transform for MODIS multispectral images

作　　者：邓黎[1] 李元祥[2] 

机构地区：[1]解放军理工大学气象学院,江苏南京211101 [2]上海交通大学空天科学技术研究所,上海200030

出　　处：《解放军理工大学学报（自然科学版）》2007年第1期99-102,共4页Journal of PLA University of Science and Technology(Natural Science Edition)

摘　　要：为了解决MODIS多光谱图像数据的存储和传输问题,提出了一种基于整数小波变换和结合波段相关性判定的谱间预测的算法,对MODIS多光谱图像数据进行无损压缩。通过整数小波变换和波段相关性判定,有效地去除了MODIS图像数据的空间和谱间冗余,其压缩效果优于当今最流行的WinRAR、WinZip压缩软件,与基于三维预测(3D-DPCM)的方法相比,该方法也能得到较好的效果,有较高的压缩比。In order to solve the storage and transmission problems of MODIS multispectral images, a lossless compression method based on integer wavelet transform and correlation estimation between spectrums was presented. The intra-band and inter-band correlations were well eliminated by the integer wavelet and correlation estimation between spectrums. In comparison with WinRAR, WinZip and 3D-DPCM, experimental results show that this method can obtain higher compression ratio.

关 键 词：整数小波变换 无损压缩 图像熵 MODIS多光谱图像 

分 类 号：TP751.1[自动化与计算机技术—检测技术与自动化装置]