基于最小包围面积矩形的密封条截面轮廓缺陷检测方法  

Sealing Strip Cross Section Contour Defects Detection Method Based on Surrounded by Minimum Area of Rectangle

作　　者：陈丽娟[1] 贺福强[1] 何磊[1] 

机构地区：[1]贵州大学机械学院,贵阳550025

出　　处：《计算机测量与控制》2014年第2期534-535,538,共3页Computer Measurement &Control

基　　金：贵州省科技计划工业攻关项目(黔科合GY字〔2011〕3037)

摘　　要：针对汽车密封条截面轮廓缺陷检测的技术需求,采用机器视觉检测方法获取截面轮廓形状,提出了基于最小包围面积矩形和基于角度直方图的联合配准方法,然后提取配准图像骨架,利用其公差带特征值判断缺陷,并通过实验验证理论方法的可行性和系统的有效性;试验证明,检测正确率平均值为93.5%,略高于人工检测的正确率90%。In automotive sealing strip cross section profile defect detection technology needs, using machine vision detection method for cross section contour shape, is proposed based on minimum surrounded area of the rectangle and the joint Angle histogram matching method, and then extract the registration image skeleton, using its value to judge defect tolerance zone characteristics, and through experiment verify the validity of the theoretical method and the system is feasible. Experiment proves that the detection average accuracy is 93. 5%, slightly higher than that of artificial detection accuracy by 90%.

关 键 词：汽车密封条 最小包围面积矩形 角度直方图 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]