Research on Metal Surface Defect Identification Method Based on HOG-LBP
Metal surface defects are complex and diverse.Efficient and high-precision surface defect identifi-cation method is the key to improve the production efficiency of metal products.A metal surface defect rec-ognition model based on HOG-LBP is constructed.The model uses median filtering to denoise metal sur-face defect images,uses gradient histogram and local binary mode to extract the surface defect image fea-tures,and uses different classifiers to make fusion decisions to achieve the goal of identifying metal surface defects.The proposed method is applied to the surface defect identification of engine camshaft,which can effectively identify the stains,scratches and dents on the camshaft surface.By comparing with other scholars'methods,it is verified that the proposed metal surface defect identification method has higher ac-curacy and efficiency,and can be effectively applied to different types of surface defect recognition.
deep learningmetal surface defectimage feature extractionclassifier
万方数据
维普
