Using Computer Vision to Recognize Defects on the Surface of Hot-rolled Steel

The problem of recognizing defects on the surface of hot-rolled steel is quite old, but technology has only recently reached a sufficient level to enable automation of this process. One of the most suitable methods is applying convolutional neural networks (CNN). We selected the Northeastern University surface defect database, which is a dataset of the most identified cases of hot rolled defects, as the qualitative dataset for network training. This article presents CNN models to recognize 6 defects with an accuracy of 93.59% and 6 defects and images of a clean surface with an accuracy of 92.31%. The recognition time was 0.001384±5% seconds for all samples. As well, program give recommendations based on the most common defects of a particular type.

I. Introduction

The task of recognizing defects on the surface of hot-rolled metal lay for a long time on the technologist, but at present this task can also be solved by computers [1–3]. The most common method is artificial neural networks [4–7]. Artificial intelligence (AI) has made great breakthroughs in recent years [8, 9]. One of the most important uses of computer vision in manufacturing is for automating quality inspection during the production process. Maintaining quality standards is of utmost importance in the field of manufacturing. While one can do this manually through engaging quality control experts, the chances of human error are quite high and naturally limited. [10, 11]. Along with the growing interest in technology and the growth of the market, companies are paying more attention to the technological advances that artificial intelligence offers. According to the study, computer vision is one of the most widely used technologies.