1.内燃机可靠性国家重点实验室，山东 潍坊 261061；2.潍柴动力股份有限公司，山东 潍坊 261061
Abstract：In order to improve the recognition efficiency of casting surface defects, the application of machine vision in engine casting surface defect detection is studied. The characteristics of casting surface defects, image denoising and segmentation technology are studied. The traditional algorithm is improved and optimized. The median filter is used to replace the mean filter for image denoising, and the threshold method is used for image segmentation; according to the system framework design, select appropriate equipment, design detection process and develop detection software. The accuracy of the surface defect detection method of machine vision casting is verified by the actual detection of casting samples. The results show that the accuracy of the detection of air hole and sand hole defects is 93.5% and 95.3% respectively. Compared with the traditional detection methods, the machine vision detection method for surface defects of castings has high accuracy, good practicability and economy, and can be better applied to the production site.
Keywords:casting surface defect; machine vision detection; image processing; edge detection; image segmentation