An experimental teaching platform for machine vision-based real-time detection taking quality inspection of paper cups as instance

In order to improve the quality of students' practical training, a real-time detection system experimental platform based on machine vision has been designed. The system has functions of real-time sampling, detection, defective product removal, result display and data storage, and its structure covers three parts: hardware platform, detection algorithm, and software function implementation. The hardware platform includes image acquisition module, material transportation module and defective product removal module, and the control circuit is based on the development of Arduino and STM32. The detection algorithm is developed based on OPENCV, and it is designed to detect three types of defects: cup rim concavity, cracking, and out-of-roundness. The software system is jointly developed using Visual Studio 2019 and Qt Creator 4.9.1, and it has good interactivity. With this experimental platform, teaching work such as image acquisition system construction, dynamic sampling problem analysis, image detection algorithm development, and interactive interface implementation can be carried out, so as to improve students' engineering design ability in many aspects.