Authors: Huang-Kuang Kung; Yan-Chen Lin; Fang-Chen Yang; Jia-Xian Chen; Yu-Li Chen
Company/Institution: Cheng Shiu University; Cheng Shiu University Mechanical Engineering; Cheng Shiu Department of Industrial Management; Cheng Shiu University Mechatronics Institute
Country: Taiwan
e-mail: eddie@wiipa.org.tw
web: https://www.wiipa.org.tw/
This research mainly uses AI deep learning to develop the technology platform for fastener internal thread defect detection. The system platform software uses AI deep learning to determine and detect whether the fastener internal thread has defects. The hardware part of the machine vision system integrates a robotic arm with 6-axis degrees of freedom. It cooperates with the imaging system to achieve internal thread image capture and defect detection in each degree of freedom. AI deep learning method is used to train the detection model to differentiate the difference between OK and NG for the internal thread of fastener. A home-made inspection user interface (UI) is developed to complete the hole thread inspection of a complex work piece as indicate in the figure. The relative hole positions and coordinates between work piece and robotic arm can be determined through a 3D modeling software. By using the home-made inspection UI and the AI-trained model, it is possible to detect internal thread defects while capturing images of the internal holes of the work piece.