Micro-defects detection on metal screw surfaces based on algorithm of faster region convolutional neural network over internet of things / Nur Aainaa Zainal

It is significant for most of the production process to develop efficient techniques in order to control products outcome. This is to ensure that the quality assurance of the products is reliable. The detection of defects in a product is one of the major production processes for quality control. The quality control process of metal screws uses much manpower for manual inspection at the manufacturing line. Manually inspecting screws of various sizes manufactured in large quantities is time-consuming. Therefore, this research proposes deep learning by implementing of Faster Region-based Convolutional Neural Network (Faster R-CNN) model for the micro defect detection on metal screw surfaces. In the meanwhile, the Internet of Things (IoT) has been identified as a suitable instrument for connectivity that enhances industrial operations with real-time monitoring; capable to provide data processing to control production quality. In this project, the defects that are considered are surface damage screw, stripped screw, and surface dirty screw. Webcam on laptop provide image in real-time is used for image acquisition of the metal screws with different types of defects. Then, the image collected is employed to train the Faster R-CNN. This programming is employed to communicate with Node-RED; as a visual tool designed for the Internet of Things (IoT) Network. The results of the experiment show that the detection accuracy of the model is 98.8%. The model also shows the superiority of Faster Region based on Convolutional Neural Networks (Faster R-CNN) in detection methods when compared with traditional machine vision techniques and Single Shot Detector (SSD Detector) model. The success of this research project in classifying the micro defect on metal screw surfaces facilitates the implementation of Industrial Revolution 4.0 (IR4.0) by the government in the manufacturing industry.