基于小波变换及Matlab的遥感图像压缩效果分析

Remote Sensing Image Compression Analysis Based on Wavelet Transform and Matlab

主要对小波变换在遥感图像压缩中的应用进行了研究。首先对小波变换理论及其实现进行了简要的介绍。然后,重点分析了小波变换分解层次、阈值选择对图像压缩效果的影响。利用MATLAB小波工具箱提供的小波变换函数,采用haar小波、db小波、sym小波、coif小波、bior小波进行遥感图像压缩,并对压缩效果进行了分析比较。结果表明,随着压缩倍数的增加,所需小波变换的分解层次也相应增加;但当压缩倍数选定后,对于大小一定的遥感图像,小波变换分解层次达到某一定级次时,再增加变换的分解层次,对压缩效果的作用甚微。因此,当图像大小一定时,压缩倍数和小波变换的分解层次之间存在一个较佳匹配问题。同时,通过实验数据,说明了在小波压缩过程中,阈值对压缩图像的压缩比和能量保持比的定量影响,表明将小波变换用于遥感图像压缩具有压缩比大,压缩效率高的优点。实验结果为小波变换在遥感图像压缩中的实际应用提供了理论参考。

Abstract In this paper, the basic principle of wavelet transform is briefly described, focusing on the influence of decomposition level and threshold of wavelet transform on emulate compression effect of RS images. The RS images are compressed by haar wavelet, db wavelet, sym wavelet, coil wavelet as well as bior wavelet which are provided by MATLAB Wavelet Toolbox. According to the analyzed and compared results, the necessary decomposition level of wavelet transform is supposed to be increased with the multiple of compression. But for the RS images with a certain size besides the chosen compression multiple, when it is increased to a certain level, the effect of compression will keep steadiness. The matching between compression multiple and the decomposition level of wavelet transform is also discussed in detail. On the other hand, the effect of the threshold to the image compression ratio and energy-keeping is studied at the same time. The experimental results and analysis indicate that high compression ratio and high efficiency satisfy the requirement on RS images compression. The result of research provides the theoretical reference of RS images compression using wavelet transform.