ABSTRACT
For high-quality products, surface fault identification is crucial. Surface defect identification on circular tubes, on the other hand, is more difficult than on flat plates due to the fact that the surface of circular tubes reflects light, resulting in overlooked faults. Surface defects on circular aluminum tubes, such as dents, bulges, foreign matter insertions, scratches, and cracks, were recognized using a unique faster region-based convolutional neural network (Faster RCNN) technique in this study. The proposed faster RCNN outperformed RCNN in terms of recognition speed and accuracy. Additionally, incorporating image enhancement into the approach improved recognition accuracy even more.
