摘要
提出一种油封缺陷检测及分类识别方法。依据量化的油封表面质量定性判断精度指标要求,结合生产实际设计并构建油封缺陷在线视觉检测系统。采用伺服电机同步传动旋转机构等分采集油封环带图像,经图像预处理分割出不同检测区域。利用小波变换模极大值图像边缘检测算法实现油封缺陷检测;提取描述缺陷的特征参量并进行主分量选择,通过支持向量机M-ary分类策略对油封缺陷进行分类识别。实验结果表明,系统及方法切实可行,具有实用推广价值。
A method for detecting and recognizing oil-seal surface defects based on image processing was put forward. According to the requirement of qualitative judging accuracy index based on quantitative oil-seal surface quality, an oil-seal surface defect online visual detection system was designed and built combining with production practice. The system adopts servo motor synchrodrived rotation mechanism to acquire the image of oil-seal aliquot circumference, and the aliquot circumference image is partitioned into different detection regions with image preprocessing algorithm. The wavelet transform modulus maxima edge detection algorithm is used to realize oil-seal defect detection;then the multiple feature values describing oil-seal defects are extracted, principal component is selected and the dimension of the defect features is reduced. Support vector machine M-ary classification strategy is used to classify and recognize the oil-seal defects. The experiment results show that the proposed system and method are feasible and have practical populization value.
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2013年第5期1093-1099,共7页
Chinese Journal of Scientific Instrument
基金
河南省教育厅科技攻关基金(201124600KP)资助项目
关键词
油封
图像处理
缺陷检测
分类识别
支持向量机
oil seal
image processing
defect detection
classification and identification
support vector machine