磨损微粒的模式识别系统的研究

磨损微粒的模式识别系统的研究 A WEAR PARTICLE RECOGNITION SYSTEM 段祥 Duan Xiang 1 2 3 first-author 胡正仪 Hu Zhengyi 1 2 3 苏祥芳 Su Xiangfang 1 2 3 宫效红 Gong Xiaohong 1 2 3 王延平 Wang Yanping 1 2 3 严新平 1 2 3 武汉大学电子信息学学院,武汉 武汉大学电子信息学学院,武汉 College of Electronic information,Wuhan University,Wuhan 430072,China)Yan Xinping College of Electronic information,Wuhan University,Wuhan 430072,China)Yan Xinping ,430063 ,430063 武汉交通科技大学船机系,武汉 武汉交通科技大学船机系,武汉 提出了一个关于磨损微粒的模式识别系统．该系统先把磨粒从背景中分割出来，然后从3个方面对磨粒进行特征提取：用于辨别磨粒的材料的特征矢量，用于辨别磨粒的轮廓形状的特征矢量，用于辨别磨粒的表面纹理特征的矢量．系统最后由这3个特征对磨粒所属的类型进行了计算机分类． Wear particle recognition is the key point in ferrographic technology. In this paper,a systemo f pattern recognition about wear particleswas proposed,which firstly segmented the wear particle frombackground,then extracted three kinds of features of the particle:a feature vector for identifying the material of the particle,a feature vector for the contour of the particle and a feature vector for surface texture.The system used three classificatorscorresponding to the three vectors to determine the particle’s featuresin these three aspects. Finally the class of the particle could be generalized from theresults of these threeclassificators. 磨粒 分割 特征矢量 分类器 wearparticle segmentation featurevector classificator TP391.41 国家自然科学基金(61672102) 1998-05-01 2021-04-01 5