Research On Key Technology Of Surface Defect Detection In Aluminum/Copper Strip

The surface defect detection of aluminum/copper strip has always been an important research direction in the field of precision inspection.With the rapid development of China’s non-ferrous metal industry,the process equipment of aluminum/copper strip production line has reached the international advanced level.However,in the fields of aviation,automobiles,large-scale integrated circuits,etc.,the quality and stability of high-end strip products still can not meet the requirements of national major projects and emerging industries.Therefore,improving the surface detection technology of high-end sheet metal is a difficult problem in the production of modern enterprises.At present,the on-line detection of surface defects of domestic aluminum/copper strips is mainly carried out by manual inspection.Manual online testing is not only labor intensive but also inefficient.In response to this problem,this paper uses machine vision and image processing technology to develop an aluminum/copper strip surface defect detection system based on the actual production workshop of aluminum/copper strip.In this paper,the surface defect detection of aluminum/copper strip is taken as the research object,and the obtained surface defect image of off-line aluminum/copper strip is studied in depth.According to the existing problems of image surface defect recognition,an improved algorithm is proposed and the correctness of the algorithm is verified.The specific work is as follows:1.Firstly,the overall scheme of the surface defect detection system is designed,including the hardware composition and algorithm improvement,and the main technical indicators of the system are proposed.2.Independently setting up the aluminum/copper strip surface defect image acquisition platform to realize real-time image acquisition of the strip surface,and the image data is transmitted to PC by the image acquisition card to realize the digital processing of the image.3.The gray level correction,filtering and denoising are used to preprocess the obtained image,and then the segmentation algorithm is used to analyze the thresholds,and an adaptive threshold algorithm is proposed.Comparing the various classical edge detection operators,the Canny operator is the best.For the binarized image,further mathematical morphology detection is performed to remove the void or virtual edge.4.Based on the analysis of traditional BP algorithm,a neural network classifier is designed,and the traditional neural network recognition algorithm is improved,and the algorithm is successfully applied to defect classification.Experiments show that the improved algorithm has faster convergence speed than the traditional BP recognition algorithm.Moreover,the performance of defect detection system for aluminium/copper strip is stable,the algorithm is simple,which can meet the performance requirements of continuous detection.