As the core of the rolling bearing components, the quality of the rolling element determines the performance of the bearing. In view of the on-line detection demand of bearing rolling element defects, the key technology of detection is studied. In view of the characteristics of the bearing image, image filtering、gray enhancement and binaryzation were preferred to complete image preprocessing. A circle edge detection method of combining annular region search and least square circle fitting was proposed to complete the detec-tion and extraction of the ROI area. The characteristics ( roundness and Area) of connected area were ana-lyzed to realize the defect detection and defect classification of the bearing rolling element. Experimental work has proved that the methods described in this paper is good real-time and high accuracy.%作为滚动轴承的核心元件,滚动体的质量决定了轴承的性能. 针对轴承滚动体缺陷在线检测的需求,对检测的关键技术进行了研究. 针对轴承图像的特点,在对图像进行滤波、灰度增强和二值化的基础上,采用了一种基于环形区域的边缘搜索和最小二乘圆拟合相结合的圆检测法,完成了轴承ROI区域的检测和提取. 为了实现轴承滚动体的缺陷检测和缺陷分类,对连通区域(圆度和面积)的特征进行了分析. 实验结果表明,所采用的方法对轴承滚动体缺陷的检测实时性好,准确率可以达到100%,具有较高的应用价值.
展开▼
