Non-destructive testing method of vehicle body weld defects based on yolov5 algorithm

In order to effectively improve the accuracy of non-destructive testing of vehicle body weld defects, a non-destructive testing method of vehicle body weld defects based on yolov5 algorithm is proposed. The image information was collected to extract the weld area, and the spatial enhancement method and median filtering method were combined to denoise the extracted weld image. After the weld defect target is detected by the combination of yolov5 algorithm and support vector machine, the improved support vector machine completes the classification and recognition of the defect category, and realises the non-destructive detection of vehicle body weld defects. The results show that the uniformity of the proposed method is maintained above 0.96, and the peak signal-to-noise ratio of the image is above 40 dB, The Pratt quality factor is always stable above 0.93, and the maximum error rate is less than 1%, which shows that the proposed method has strong detection performance.