图像中边缘信号高斯平滑后的偏移问题

摘要

Gauss smooth is a common pretreatment way, but it can't eliminate noise in small scale and cause image edge offset in large scale. The paper analyzed it,proposed edge offset compensation method under lager scale gauss smooth, and solved the problem of gaining precise and less noise image edge by using multi-scale theory.%高斯平滑是一种常见的预处理方式,但小尺度下的高斯平滑不能完全消除噪声,而大尺度下的高斯平滑可能引起图像边缘的偏移.对此进行了分析,提出了大尺度高斯平滑下的边缘偏移补偿方法,并利用多尺度理论解决了下获得图像的精确的、少噪声的边缘难题.