基于预测匹配差与全局-局部阈值化的轴承缺陷检测与定位算法

Bearing Defect Detection-Location Algorithm Based on Predictive Matching Difference and Global Local Thresholding

为了准确检测轴承在生产加工过程中出现的滚动体漏装等缺陷,提出了基于预测匹配差与全局-局部阈值化的轴承缺陷检测与定位算法,完成滚动体缺失、破损检测与定位。首先,引入分段线性图像增强技术,扩大滚动体与轴承背景的对比度;其次,综合全局与局部阈值化方法,结合种子填充技术,对轴承进行连通边缘标记;再设计一种圆验证机制,将轴承中的非圆边缘滤除,以提取滚动体的ROI区域,缩小了目标检测范围,提高滚动体缺陷的检测效率;最后,利用Open CV来统计不同部件的轮廓面积,从而设计预测匹配差方法,对缺失或破损的滚动体进行定位。仿真结果显示,与当前轴承检测方法相比,对于滚动体漏装或破损轴承,所提算法具有更高的检测与定位准确。

In order to accurately detect the defects of rolling body leakage in the process of production bearing,the bearing defect detection-location algorithm based on predictive matching difference and global local thresholding w as proposed in this paper to finish the detection and location of missing and damaged rolling bodies. Firstly,the piecew ise linear image enhancement technique w as introduced to enlarge the contrast betw een the rolling body and the bearing. Then the linking edge marking of the bearing w as done by constructing the global-local thresholding method and combining the seed filling technology. And the ROI region of the rolling body w as extracted by designing a circle verification mechanism to filter out the non circle edge of the bearing for reducing the target detection range and improving the detection efficiency. Finally,the predictive matching method w as designed by using the Open CV to calculate the contour area of different components for locating the missing or damaged rolling objects. Simulation results show that this algorithm has higher detection and localization accuracy compared w ith current bearing detection methods.