基于遗传算法和矩的刀具磨损图象检测技术的应用

研究了在遗传算法基本原理的基础上 ,用最大类间方差法对刀具磨损图象进行阈值分割 ,然后对分割后的二值刀具磨损图象 ,利用一种新的基于边界的几何矩的快速分割算法进行几何矩的计算和分析。通过试验 ,研究了经过遗传算法进行优化阈值分割后的二值图象的几何矩的值的大小和刀具磨损状态的关系 ,结果表明 ,几何矩的值的大小和刀具的磨损状态密切相关 ,几何矩的值的变化率可以对刀具状态进行识别 ,对试验数据的测试表明 ,这种方法能有效提高刀具磨损状态识别的正确率

The method of tool condition monitoring based on GA and moments

A novel algorithm for adaptive threshold selection from the abrasion image of tool using genetic algorithm and geometric moments is presented. The principle of threshold selection is to use maximum classes square error as an example to describe the method and procedure about applying genetic algorithms in optimizing the selection of image threshold, then use a new fast calculation method of geometric moments for the segmented binary tool image, which is calculated through two phases. According to this method, the relationship between the value of the geometric moments of the binary tool image and the tool conditions was investigated. The results indicate that the value of the geometric moments of the binary tool image closely correlates with the wearing state of tools. The sensitivity of the value of the geometric moments expressing state of tool is very high.