Defect Detection and Classification Using Machine Learning Classifier

CLASSIFIER Mitesh Popat and S V Barai 1 Johns Hopkins University, Baltimore, USA 2 Indian Institute of Technology, Kharagpur, India. Abstract: In most cases visual inspection of the hot strip by an inspector (in real time or videotaped) is a difficult task. The issues in this project study are data modeling, Machine Learning (ML) model neural networks (NN) modeling and reliability of such models for automatic detection and classification of defects of hot strips. The proposed study intends to develop general guidelines for developing NN model for automatic surface inspection for hot strip mills. Introduction: In steel industry, visual inspection of the hot strip by an inspector is, in most cases not possible because of the high speeds and high temperature involved. In recent times, only video monitors and video recorders have been used where inspectors check on-line or taped video sequences for defects. In this way, only small parts of the strip’s top or bottom side are viewed. Additionally, this visual inspection is subjective and dependent on a large number of human factors such as the problems of working through night shifts, attention being drawn to other events, subjectivity in terms of defect severness assessment and restricted capabilities to cover the entire strip at high line speeds. Developing automatic detection and classification of surface defects of hot strips has been really a challenging problem in the field of steel manufacturing industry (Rinn et al., 2002). An automatic detection and classification system requires knowledge of data concerning the current state of the hot strips, and a methodology to integrate various types of information into decision-making process of evaluating the quality of the product. The need for surface detection technologies for surface defect classification has long been recognized. Real time image processing typically involves the application of high-speed camera, which may give the defect images in real time. The large amount of information is gathered during this process in the form of images. Human experts evaluate this information and give the level of defects, types of defects of the hot strips and will suggest remedial measure to over come those defects. However, the complexity of such program could be avoided by creating a Decision Support System (DSS). This study will take the initial steps in developing a DSS for hot strip evaluation in the manufacturing plant based on the process carried out in the practice. Typical paradigm is shown in Figure 1.