Research On The Analysis System Of Weld Surface Defects Of Tank Trucks Based On Machine Vision

As a kind of atmospheric metal tanker for transporting hazardous chemicals,tanker trucks need to be inspected regularly to eliminate potential safety hazards due to the particularity of their transport media.At present,the inspection of tank trucks often uses manual sampling and measurement.Long-term inspection in the tank seriously affects the health of the inspectors.The internal welding seam of the tank truck is most prone to defects due to stress concentration,and this part happens to be the most difficult to detect.In this paper,machine vision method is used to analyze the welding seam defects on the inner wall of tank trucks,and a welding seam extraction method based on edge detection is proposed.The main component analysis and support vector machine methods are used to realize defect classification.Finally,a tank truck based on human-computer interaction interface is designed.The welding seam surface defect analysis system realizes the digital and intelligent detection of tank truck defects,which greatly reduces the inspection time in the tank and guarantees the life safety of the inspectors.The specific work is as follows:First,the requirements of the tank truck detection system are analyzed,and the preprocessing method of the tank truck weld image is studied.The weighted average method is used to gray image,and the denoising effect of various filtering denoising methods on weld gray image is studied.Aiming at the problem of edge blur after filtering,the edge of the weld is enhanced by analyzing the change rate of image pixels.Secondly,the method of extracting welding seam defects of tank trucks is studied,and the welding seam defects are extracted.Aiming at the problem of low image contrast of tank truck welding seam defects and uneven image illumination,this paper proposes a welding seam extraction method based on edge detection,which uses edge detection,morphological processing,area threshold filtering and minimum bounding rectangle to extract parallel Welding seam,using the "rotation method" to extract the inclined welding seam.On this basis,the threshold segmentation method is used to segment the weld defects,and the feature selection method is used to remove the noise around the weld defects,so as to realize the extraction of weld defects.Then,the welding seam defects of the tank car are analyzed,and the welding seam defect classification is realized.Ten features related to weld defects are selected,and principal component analysis is used to reduce the ten-dimensional defect features to three-dimensional,with a cumulative contribution rate of 92%.The support vector machine method is used to analyze the defect feature data after dimensionality reduction,and the average accuracy of defect classification is 99.7%.Finally,a surface defect analysis system for tank truck welds based on a human-computer interaction interface is designed to realize the functions of image collection,defect extraction,defect analysis and image management.