刮板输送机中部槽焊缝缺陷机器视觉识别  

Machine Vision-based Defect Recognition Processing System of Welding Line on Central Trough of Scraper Conveyor

作　　者：孙昌立[1] 张剑飞[1] 

机构地区：[1]黑龙江科技学院计算机与信息工程学院,150027

出　　处：《中国科技信息》2009年第9期25-25,42,共2页China Science and Technology Information

摘　　要：焊缝透视的数字化图像由CCD摄像机拍摄焊缝胶片来获取,对图像进行中值滤波和小波变换等去噪处理,分割图像用熵关联法,roberts确定缺陷边缘,提取缺陷,分析焊缝缺陷特征,选取识别参数,建立了识别焊缝缺陷的支持向量机模型。实验结果表明该方法识别率高、稳定性良好。Digital image of welding line film was got by CCD sensor, which is preprocessed by wavelet, median filter, fuzzy enhancement. The entropy method was used to segment the image, and Roberts to detect the edge and extract the defect. Defect features were analyzed and parameters selected, then SVM （support vector machine）model was developed. The experiment results show that the system has much better properties.

关 键 词：CCD摄像机 焊缝缺陷 图像处理 识别 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]