A Defect Recognition Method for Auto Parts Based on Similarity Matching

Abstract

The invention relates to an auto part defect recognition method based on similarity match. The method has functions of detecting a leakage welding point, a leakage small part, a leakage nut and a leakage punching hole on the auto parts, and the detection objects are a welding point, a small part, a nut and a punching hole. The method comprises realization steps: firstly, an auto part image is acquired, and whether a template library needs to be built is judged; then, a detection area is captured on the acquired image, and image preprocessing is carried out on the area; then, template images ina detection object template library are traversed, the template images and the image after pre-processing are subjected to scanning and similarity contrast, a contrast result is saved to a cue map, and all information of the detection objects is included in the cue map; and finally, contour detection processing is carried out on the cue map, a contour perimeter, a roundness and an area are used to select a detection object, the contour number and position information is acquired, and whether defects of a leakage welding point, a leakage small part, a leakage nut or a leakage punching hole exist on the auto part in the acquired image can be detected.

Description

technical field [0001] The invention relates to image processing technology, in particular to a method for identifying defects of auto parts based on similarity matching. Background technique [0002] A car is composed of tens of thousands of parts, including body-in-white assembly parts and plastic parts, which account for a large proportion of the car. Its OEMs have very high requirements for the quality of its products. The parts of the distribution manufacturers require that the pass rate of key parts reach 100%, and other parts basically have to reach 99.99%. For first-tier suppliers, the requirement is within 100ppm. Other secondary and tertiary suppliers are 200-400ppm. Therefore, whether the parts are qualified or not directly affects the quality of the OEM products, and the distribution manufacturers must spend a lot of manpower to detect and troubleshoot defective parts. [0003] From the perspective of the automobile market, at present, domestic small and medium...

AI Technical Summary

Problems solved by technology

[0003] From the perspective of the automobile market, at present, domestic small and medium-sized enterprises mainly rely on manual inspection, which has low work efficiency, high probability of missed inspection, and high labor costs. In view of the above-mentioned situation in the domestic automobile industry, the research and development of the parts visual inspection system is to solve the factory inspection. Problems such as low efficiency, large number of personnel, and low detection quality

Images