Research on Bridge Crack Recognition Algorithm Based on Image Processing

Abstract

Concrete is a common material of bridges. Due to durability, high pressure, and other reasons, it is easy to cause cracks on the surface of bridges, which affects the appearance and causes structural safety problems. The traditional manual testing of bridge crack is time-consuming and laborious. Therefore, we propose an algorithm for concrete crack recognition based on image processing. Images for the crack area are taken by using high-definition cameras. The collected image is pretreated for graying, denoising, and filtering with the treatment of dummy pixel and fracture problems. The improved multistage filtering algorithm was used to extract the complete crack, and the width and length of the crack were calculated. Experimental results show that the proposed algorithm effectively detects concrete surface cracks with high accuracy.