基于机器视觉的大尺寸薄壁机械零件微裂纹检测研究

Research on Micro-crack Detection of Large-size and Thin-walled Mechanical Parts Based on Machine Vision

当前的微裂纹检测方法不能对微裂纹图像进行平滑处理,导致无法有效检测到微裂纹的长度、面积以及圆度。为此,设计一种基于机器视觉的大尺寸薄壁机械零件微裂纹检测方法,对大尺寸薄壁机械零件微裂纹图像进行灰度拉伸,利用邻域均值算法对拉伸后的图像进行平滑处理,进而利用机器视觉理论提取微裂纹图像缺陷特征,通过计算对微裂纹的缺陷特征进行约束处理,完成大尺寸薄壁机械零件的微裂纹检测。测试结果表明:该方法具有较好的检测效果和精准度。

To overcome the poor effective detection of the length,area and roundness of micro-crack due to its unsmooth image worked by current detection methods,a micro-crack detection method for large scale thin-walled mechanical parts based on mechanical vision is designed,with which the micro-crack image of large scale thin-walled mechanical parts is gray stretched,the stretched image is smoothed by the neighborhood mean algorithm,the defect characteristics of micro-crack image are extracted by machine vision theory and the defect characteristics of micro-cracks are constrained by calculation so as to complete the micro-crack detection of large size thin-walled mechanical parts.The test results show that the designed method has good detection effect and accuracy.