用最小误差确定浓淡图象的阈值

A METHOD FOR GRAY-LEVEL PICTURE THRESHOLDING USING THE LEAST ERROR PRINCIPLES

在提出和论证最小误差原理后,通过测量参数自动确定阈值的数目,使图象处理中多阈值选择问题得以解决。用最小误差确定浓淡图象的阈值是一种自适应无监督的图象分割方法,根据文中算法可以确定各种形式直方图的阈值。

Threshold selection is one of the most important problems in picture processing and pattern recognition. In this paper a least error theory is presented. By this method the number of threshold can be determined automatically and problems of multithreshold selection can be solved. It is a self-adaptive method and need not supervise image segmentation. With the proposed algorithm thresholds for any gray histogram can be selected.