灰度图象的最大最小优化熵阀值分割

The Maxmin OptimaI Entropic Thresholding of Grey－level Images

最大熵阀值法是最常用的利用图象灰度直方图的阀值化方法之一，它是基于分割后物体与背景分布的信息熵为最大的优化准则来自动选择分割阀值。本文给出了另一种优化熵准则─—最大最小优化熵准则来自动选择图象分割的阀值。实验结果表明，当物体与背景的灰度分布呈现正态分布时，本文的优化熵阀值比最大熵阀值具有更好的分割性能。

Abstract：One of the most useful thresholding techniques using the grey－level his-togram of an image is the maximum entropy method．Its thresholding selection is automati-cally solved by maximizing the information entropy of the objects and background distribu－tion in the segmented images. This paper presents a new optilnal entropic method that themaximum minimum（maxmin）optimal entropic criterion automatically selects the thresholdof image segmentation. The comparison between this method and the maximum entropymethod shows that this method perforrns much better when the histogram of an image is ofnormal distribution.