基于灰色聚类的磨粒自动识别

 将灰色系统理论中的聚类分析技术用于磨损颗粒的自动识别, 编制了相应的计算机模拟程序。在对磨粒图象的形态特征参数进行敏感性分析的基础上, 确定了各参数的灰类白化权函数, 并结合磨粒识别的试验研究, 给出了磨粒各特征参数的聚类权值。应用此方法对一组标准磨粒进行了模拟识别, 识别正确率在90% 以上, 并且识别速度很快, 大大优于传统的磨粒识别方法。

Wear Particles Identification Based on Grey Clustering Analysis

As a new technique, the theory of grey system has been applied to many areas such as perform prediction, relational analysis and decision making. In this paper, the whitenization weight function of grey clustering is presented after sensitivity analysis of catachrestic parameters of wear particles, and the clustering weight is also given based on the identification test of wear particles. The program of auto identification of wear particles has been made by means of grey clustering analysis, and an experiment of wear particles classification by this method has been done. The identification accuracy of debris by grey clustering analysis is higher than 90%, and the speed of classification is very fast. It is much better than the traditional ones.