针对特征点区域无损压缩进行编码,通过编码标志点检测识别特征点区域,将特征点区域和背景区域分别进行处理,特征点区域采用整数小波变换无损压缩,背景区域采用离散余弦变换有损压缩。为了更直观显示算法的特点,将压缩图像与TIFF压缩和JPEG压缩进行对比。实验结果表明使用该算法获得的压缩图像既节省了存储空间,又将特征点区域的信息完整地保存下来。适用于视觉测量中特征点区域无损压缩的研究。
In the paper we encode for the lossless compression of feature points zone, and detect and recognise the feature points zone through coding mark targets. We process the feature points zone and the background zone separately, in feature point zone we use lossless compression of integer wavelet transform, and in background zone we use lossy compression of discrete cosine transform. In order to display the structure of the algorithm in the paper more intuitively, we compare the images compressed in the way of this paper with TIFF compression and JPEG compression. Experimental result shows that the compressed image derived from the algorithm in this paper saves the storage space while entirely preserves the information of the feature points zone. This algorithm is suitable for the research on lossless compression of feature points zone in visual measurement.