AI CNN Based Defect Inspection System of a Governor Using a Rotating Platform

In this study, an automated defect-detection system based on a convolutional neural network (CNN) for rapid governor rotation was developed and its performance was evaluated. In the proposed study, the governor was used in diesel engines to regulate the fuel injection amounts to regularly maintain the engine&quots rotational speed. The governor was rotated using a servomotor at speeds of 700, 1000, 2000, 2800, and 3200 rpm, and the degree of the governor’s opening was captured using a camera. CNN was employed to detect the defect states using captured images of the governor&quots opening degree. The experimental results compare and evaluate the performance using two CNN algorithms: Xception and Inception. This study demonstrates the application of CNN-based models to improve defect-detection systems in manufacturing processes. Accordingly, CNN algorithms extract and learn features from images, rendering them useful for the detection of various types of defects.