现代电子技术2017，Vol.40Issue(23)：47-50,4.DOI:10.16652/j.issn.1004-373x.2017.23.012
面向再制造服务的回转类表面缺陷识别
Surface defect recognition of rotational parts oriented for remanufacturing service
摘要
Abstract
For the complex environment in remanufacturing activity fields,it is difficult to acquire the clear image with sur-face defect in the remanufacturing target. Aiming at this problem,a surface detect recognition method of rotational parts is pro-posed,which is based on the characteristics of frequency domain and spatial domain. The nonsubsampled Contourlet transform is used to extract the Contourlet coefficient characteristics of original image under different dimensions,and then the characteris-tics are combined with the gray textural characteristics in the image. The PCA method based on radial basis kernel function is adopted to perform the dimension reduction for the original characteristics. The least square probability classification method is employed to classify the characteristics after dimension reduction. The case analysis is carried on for the image acquired on site. By using the proposed classification and recognition method,the average classification accuracy of the four types of the remanu-facturing roll surface defect is higher than 92%.关键词
再制造服务/表面缺陷识别/非下采样轮廓波变换/核主成分分析Key words
remanufacturing service/surface defect recognition/nonsubsampled contourlet transform/kernel principal com-ponent analysis分类
信息技术与安全科学引用本文复制引用
刘翔,夏绪辉,王蕾..面向再制造服务的回转类表面缺陷识别[J].现代电子技术,2017,40(23):47-50,4.基金项目
国家自然科学基金(71471143)资助项目 （71471143）
