【摘要】 目前通用的变换域图像压缩算法由于其工作原理导致其在压缩具有丰富纹理细节的遥感图像时质量不高,为了满足当前航天遥感技术的需要,提出了一个基于在空间域上对图像结构进行渐进描述的图像压缩算法。该算法把分块图像根据其内部的纹理复杂程度分为平坦、粗纹理和细纹理3种模式,并利用纹理简单模式叠加修正的方式来渐进地描述纹理复杂模式。在压缩细节纹理较少的图像时,该算法的结果与基于小波变换的典型算法SPIHT相当,而在压缩纹理细节丰富的图像时,该算法明显优于SPIHT算法。更多还原

【Abstract】 An image compression algorithm based on the progressive description of spatial structure was proposed.In the algorithm,each subimage was assorted into one of the three patterns: the smooth area,the coarse texture area and the fine texture area according to the texture complexity of the subimage,and the more complex pattern was made from the superposition of the less complex pattern and the related correction.In comparing with the typical wavelet based image compression algorithm SPIHT,the compression results show that,when compressing images with less details,the results of the two algorithms are about the same,but when compressing images with rich details the new one are superior to SPIHT.更多还原