改进的矩不变阈值分割图像算法

The Improved Algorithms of Thresholding of Images by Keeping the Intensity Moment Inbariable

介绍图像的矩以及在保持图像的矩不变的条件下对图像进行两级阈值分割的方法,指出利用该方法分割出来的图像具有过多的无用背景信息.为此,在矩不变阈值算法的基础上提出了两种改进算法,一种是根据特征点检测函数进行迭代;另一种是根据特征点检测函数的均值进行调整,以达到尽量保留图像前景物体信息,去掉无用背景信息的目的.

This article introduces the intensity moment of images and one method used for the bi level threshol- ding of images under the condition of keeping the intensity moment invariable. But the images segmented have much useless background information, so we put torward two improved algorithm~ based on this method. One is to use iteration and the other is to use mean adjusting according to the inspective function of points of interest. The two methods can hold the useful foreground information and get rid of the useless background information.