Stereoscopic vision on-line detection method for surface crack of stretched aluminum alloy plate

Abstract

The invention discloses a stereoscopic vision on-line detection method for a surface crack of a stretched aluminum alloy plate. According to the vision detection method, two CCD (Charge Coupled Device) cameras are fixedly connected with equipment so as to keep the constant parameters of the relative position of the CCD cameras; a matched calibrating program is used for automatically finishing a calibrating operation for the parameters of the two cameras; and a special precise mechanical system is used for controlling and outputting focal length parameters of the cameras, so as to realize the precise measurement for space coordinate information of a small object. Accordingly, a real-time measuring algorithm for an area stress and a crack length on the basis of displacement of a surface texture feature point (detecting point) is provided. The measured stress level and crack length information are combined with a crack growth theory, so that the crack threatening estimation and the action forecasting are realized, the optimal stretching technical parameter is confirmed, the belt breaking accident is avoided during a stretching process, the safety of equipment and production personnel is ensured, the yield of products is increased and the application prospect in the panel processing field is wide.

Description

technical field [0001] The invention relates to a plate stress and crack geometric dimension detection technology, in particular to a computer vision-based on-line detection method for surface cracks of a stretched aluminum plate. Background technique [0002] In the sheet and strip drawing process, the quality defects of the sheet to be processed, the clamping conditions, the drawing process parameters and other factors will cause the surface cracks of the sheet and strip. It not only determines the quality of the finished product, but may also lead to serious belt breakage accidents, damage the stretching equipment, and threaten the personal safety of the production personnel. Therefore, it is necessary to strictly monitor the development of surface cracks in real time, adjust the process parameters in time, and ensure the production quality. went well. [0003] At present, the detection of surface cracks is mainly divided into two methods: active and passive. The active...

AI Technical Summary

Problems solved by technology

Under production conditions, the vibration and electromagnetic interference of the processing equipment itself will be coupled into the excitation signal and data acquisition signal of the active detection system through various channels, which can significantly reduce the accuracy of crack measurement, and may not be able to detect small cracks; At present, the passive detection technology mainly detects cracks by using image processing and recognition methods such as wavelet analysis and neural network, which is more focused on solving the problem of "existence" or "absence" of cracks in a certain area, and cannot give the geometric size information of cracks. And this is precisely the key basis for adjusting the stretching process parameters

Images