Defects Detection and Classification on Metallic Surfaces

Inspecting metallic parts is a challenging task and is in demand in various manufacturing quality control applications. Computer Visions (CV), Machine Learning (ML) and Deep Learning (DL) are typically used for automating the inspection process, in particular, for detection, recognition, and classification of metallic surface imperfections. The objectives of this research are bifold: 1) programming a robotic arm to pick and drop objects and capture multiple images of the object using a stationary camera and 2) developing the CV, ML, and DL algorithms for identifying defects such as dents, scratches, glue and paint marks on smooth metal surfaces.