Visual inspection method of steel pipe surface cracks based on dry magnetic particle feature enhancement

Abstract

ABSTRACT It’s challenging to use the machine vision to inspect the hot-rolled steel pipes with complex surface texture, uneven colour greyscale, and micro-cracks. On the basis of the principle of magnetic particle detection, a visual inspection method of steel pipe surface cracks based on the dry magnetic particle feature enhancement is proposed, in which the dry red magnetic particles are applied to the steel pipe’s surface to form magnetic particle indication, and complementary colour images are captured in red and green light sources, and image difference is used to eliminate background noise and enhance magnetic particle indication features in view of the fact that the red light source improves and the green light source reduces the magnetic particle indication greyscale. Afterwards, the principle of selection of complementary colours and the detection principle are presented and analysed. Then, the magnetic particle images acquisition and the processing operations of the magnetic particle indication feature enhancement and crack segmentation are presented in detail. Finally, the feasibility of the method proposed in this paper is verified by experiments, and image segmentation metrics demonstrate that it has a good detection effect on the hot-rolled steel pipes with complex surface texture, uneven colour greyscale, and fuzzy cracks.