Software Development Of Defect Detection System For Die-cutting Product Based On Machine Vision

The die-cutting product is the spare part formed by cutting and precision processing the raw materials in a predetermined shape.With the rapid development of the consumer electronics industry,the massive demand for die-cutting product makes the quality control in the production process particularly important.The traditional manual offline defect-based spot inspection method has problems such as high cost,insufficient inspection rate,and easy introduction of additional damage.Therefore,it is of great value in engineering application to develop an automatic high-precision on-line defect detection system for die-cutting product.In this thesis,according to the requirements for the automation,real-time,correctness,and traceability,the software of defect detection system for die-cutting product based on machine vision has been developed.The system implements functions such as defect detection,real-time detection information push,and historical inspection data traceback and analysis.The software of the system includes online inspection software,data service platform software and remote monitoring software.The online inspection software is responsible for the execution of the defect inspection task,it supports the acquisition of relevant parameters from the DXF format files,and uses the shared memory to transmit real-time inspection information to the data service platform.The data service platform software supports message interaction,it can respond to remote requests and push real-time detection information to the monitoring center.The remote monitoring software supports unified management and configuration of multiple defect detection sites,and it can also display,store and analyze the received real-time detection information.Also,this thesis has developed a kind of defect detection algorithm,and used the embedded GPU and CUDA to achieve parallel optimization of the algorithm.After testing,the system achieves accurate and efficient defects detection,also supports remote monitoring,backtracking and analysis of data.The deviation of the system's dimensional defect detection is no more than 0.02mm,the surface defect detection accuracy rate is over 97%,and it can ensure real-time online detection of 8 meters per minute.It satisfies the system's functional and performance requirements.