Date of Publication :27th December 2022
Abstract: Inspection of metallic parts is a challenging task and has demands in a variety of manufacturing quality control applications. Computer visions are typically used for the purpose of inspection, in particular, for detection, recognition, and classification of surface features representing manufacturing imperfections. This research focuses on automating the process of defects detection using Computer Vision (CV) and Machine Learning (ML) algorithms. The objective of this research is bifold: 1) programming a robotic arm for capturing multiple images of the object and 2) developing CV and ML algorithms for detecting and classifying defects. This paper discusses different types of surface defects on metallic objects. This paper also compares the performance of EfficientDet and a ResNet neural networks model for surface defects detection.
Reference :