广东石油化工学院学报Issue(3)：30-34,5.
基于支持向量机的铁谱磨粒自动识别
Ferrograghy Abrasive Automatic Recognition Based on Support Vector Machine
摘要
Abstract
In order for the support vector machine to automatically recognize abrasive ,firstly ,color abrasive images are processed ,success-fully segmenting abrasive from the image by using K-means clustering ,region growing method and mathematical morphology .Secondly , the shape and size ,texture and color characteristic parameters are determined according to morphologic characteristics of abrasive ,and ap-propriate methods are used to extract the parameters of these three areas .The different support vector machine parameters have a great im-pact on its classification ,so genetic algorithm is used to optimize the parameters further .Finally ,the optimized SVM is used to identify sev-er sliding ,spherical ,cutting ,fatigue and red oxide abrasive ,and the recognition accuracy rate achieves 90% .The results show that this method is feasible .关键词
支持向量机/自动识别/K-均值聚类/遗传算法Key words
Support vector machine/Automatic recognition/K-means clustering/Genetic algorithm分类
机械制造引用本文复制引用
邱丽娟,宣征南,张兴芳,何照荣,孙志伟..基于支持向量机的铁谱磨粒自动识别[J].广东石油化工学院学报,2015,(3):30-34,5.基金项目
茂名市科技计划项目（201327）；广东省石化装备故障诊断重点实验室开放基金项目 （）
