Evaluation method of body welding quality information based on difference convolutional neural network

Abstract

Evaluation method of car body welding quality information based on difference convolutional neural network. The method includes the following steps: Step 1, vehicle body image acquisition: use a high-definition camera to collect multi-directional welding images of the vehicle body, where the orientation includes six directions: front view, top view, left view, right view, bottom view and rear view; step 2, loss Feature construction: use the proposed relative perturbation difference loss to construct the loss features of the images collected in six directions and the standard image; step 3, model offline training: use the loss features constructed in step 2 to train the difference convolutional neural network until Network convergence; step 4, model online discrimination: install the difference convolutional neural network trained in steps 1 to 3 in the host computer, and judge the quality of the collected body welding images in real time. The method of the invention can accurately and effectively judge the welding quality of the vehicle body, and has good practical application value in strictly controlling the factory quality of the automobile.

Description

technical field [0001] The invention relates to the field of vehicle body welding quality assessment, in particular to a method for judging vehicle body welding quality information based on a difference convolutional neural network. Background technique [0002] The body shell is a complex structural part, and it is also a typical welded structural part. For example, the body of a van is made of hundreds of stamping parts, projection welding nuts, projection welding bolts and other components connected by spot welding, projection welding, CO, gas shielded welding and other processes. The quality of welding is not only related to the beauty of the vehicle body, but also directly affects the quality of the entire body, and even endangers personal safety. Therefore, the quality of body welding must be strictly controlled. The purpose of welding quality management must be to reduce production costs and ensure that the quality meets the technical requirements of the product, an...

AI Technical Summary

Problems solved by technology

The invention is simple in structure, easy to operate, low in cost, and can realize real-time detection of the positioning of welding fixtures on the body production line, but the patent does not detail the evaluation method of body welding quality

Images