Background: Online inspection technology of cold-rolled steel surface defects has become an important means to improve product quality. This paper focuses on the development of cold-rolled steel strip surface defects detection system using the line light source and line scan cameras.
Objective: The image of the strip surface needs to be of sufficient quality so that defects can be detected accurately. Basically, this concerns image gray level. The image processing of defects detection needs to overcome the influence of the noise points and complete the defects contour extraction well.
Method: A closed-loop control method is studied. It evaluates the gray value of captured image and ad-justs the exposure time of line scan camera accordingly. A six steps image processing method is pro-posed. It could complete the contour extraction of surface defects well and ensure the processing time.
Results: Experimental results show that the system can realize online detection for cold-rolled steel strip surface defects.
Conclusion: The defects of cold-rolled steel strips need to be recorded. The patent proposed in this paper provides a new way of surface defects inspection based on machine vision.