Windshield scratch continuing detection device and method based on machine vision

Abstract

The invention discloses a windshield scratch continuing detection device and method based on machine vision, and the detection device comprises a conveying device, a cleaning device, a lighting device, and an image collection and processing device; the conveying device is integrally installed on a supporting frame, the upper and lower parts of the right side of the supporting frame are darkrooms,and the surface of the right side of the supporting frame is hollow. The cleaning device, the lighting device and the image collecting and processing device are arranged in the darkroom. The windshield is fixed through the conveying device and conveyed to the darkroom, the cleaning device cleans the glass entering the darkroom, and the using number of the lighting devices is determined by the sizeof the glass. The industrial cameras above and below continuously detect the front and back surfaces of the glass, and send shot images to the industrial personal computer to process and analyze theimages to finally judge whether scratches exist on the surface of the glass or not. The detection device analyzes images based on machine vision and can adapt to windshields of different sizes, the number of the lighting devices can be adjusted, and electric power is saved.

Description

technical field [0001] The invention relates to the technical field of automobile detection, in particular to a machine vision-based windshield scratch detection device and detection method. Background technique [0002] As an important part of the car body, the front windshield of the car is the first barrier to protect the safety of the driver and passengers. If the cracks in the front windshield are not repaired in time, it will have a great impact on the subsequent use of the car. Therefore, people pay more attention to the detection of scratches on the front windshield. [0003] At present, the detection of the front windshield of a car is generally done by manual observation, which not only consumes a lot of manpower, but also has limited scratches that can be observed manually, and some extremely small scratches may not be observed, which is easy resulting in inaccurate test results. The Chinese patent with the application number 201720879709.2 provides a scratch d...

AI Technical Summary

Problems solved by technology

[0003] At present, the detection of the front windshield of a car is generally done by manual observation, which not only consumes a lot of manpower, but also has limited scratches that can be observed manually, and some extremely small scratches may not be observed, which is easy result in inaccurate test results

The Chinese patent with application number 201720879709.2 provides a scratch detection device. The device drives the bearing plate and the glass on the bearing plate to rotate through the motor, which improves the detection efficiency and reduces the labor intensity of personnel. However, because the glass is a hyperboloid structure , this method can only detect one side, and is essentially manual detection, which has shortcomings

The detection scheme of the Chinese patent with the application number 201720843553.2 can adjust the height of the device and expand the scope of detection. However, the light is irradiated on the glass through the transparent plate. If there is dust or scratches on the transparent plate, the detection accuracy will be reduced. At the same time, the glass is scratched. It is still obtained through manual observation

Images