Strip Weld Detection System Based On Machine Vision

Strip steel is one of the indispensable materials in modern industry and construction industry.During the production process of strip steel,steel strip coils need to be made into coils for the convenience of subsequent transportation of strip steel.The steel strip coil is welded and rolled from the end to the end of a strip of steel.During the rolling process,it needs to go through multiple reversing rollers for reversing.In the commutation process,if the welding quality of the steel strip is unqualified,it will cause the weld to break and affect the quality of the strip.In order to ensure the qualification rate of the strip steel,it is necessary to carry out quality inspection on the strip weld seam.For the quality inspection of strip steel welds,this thesis implements a strip steel weld inspection system based on machine vision and completes the overall scheme design,realizes the cutting of weld samples and the quality inspection of weld seams.In the cutting of the strip weld sample,the industrial camera is used to collect the image information of the strip weld position,and it is preprocessed through image filtering to eliminate the interference factors and improve the image quality.The edge of the weld seam is obtained by using the amplitude method of the concave-convex line segment so as to realize the location of the weld seam.In order to solve the problem of inaccurate positioning of some weld edge points,the position maximum distribution frequency filtering algorithm is used to suppress and correct incorrectly marked points.The location algorithm is used to determine the weld position of the strip steel by using the image collected by the camera,which is used to complete the subsequent welding seam sample cutting task.Finally,different models were compared on the strip weld detection system.The recognition rate of the improved VGG16 model in the test set reached 98.1%,reaching the goal of the strip weld detection system.