SUSAN算子在微小轴承表面缺陷图像分割中的应用

Application of SUSAN operator in image segmentation of micro bearings surface defect

采用计算机视觉识别技术对微小轴承端盖上的缺陷进行了自动识别。运用现代控制理论、图像采集与处理技术、软件编程技术,针对缺陷的位置、面积和深度所具有的随机性,设计了一套适合于检测微小轴承表面缺陷的系统。该系统可以采集到高质量的图像信息。利用SUSAN算子实现了对图像的分割和微小轴承表面缺陷的快速检测。实验结果表明,该方法的识别率可达到96%,并具有方法简单、定位准确、抗噪能力强、计算速度快、实时性好等优点。

The automatic inspection system of micro bearings surface defect is studied by the technology of computer vision identification. With the help of some integrated theories, such as the modern control theory, the technology of image acquisition and processing as well as the technology of software programming, a set of system is designed, which aims at random changes of the defect position, area and depth. The system points out that it can acquire complete images information, and the micro bearings surface defect can be rapidly inspected by the way of dividing the images with the SUSAN operator. The experiment indicates that the above system has the advantages of convenience, correct location, the capability of noise resistance, high speed processing and temporal correctness. The inspection accuracy of this system can reach as high as 96 %.