Using Image Processing for Automatic Detection of Pavement Surface Distress

Abstract

The layers of asphalt pavement can develop a variety of issues and faults, each of which is caused by one or more factors (such as bad mixture design, improper paving techniques, or environmental factors), with traffic being the primary cause of the majority of them. In this paper, a crack detection technique in the pavement image is proposed to address the issue based on the edge information. To accomplish this yhe image is first pre-processed to improve the linear characteristic of the crack by converting the colorful digital image to gray scale using the gray-scale transformation function and the reconstruction filter. The adaptive thresholding method is also designed to map the crack gradient information while coarsely extracting the crack edge based on the grayscale feature. After the filtered edge points have been gathered in line with the gradient information, the edge is recognized using single-pixel filtering processing, which is enhanced by utilizing the local difference between pixels in the fixed region. The complete crack is acquired by filling the crack edge. The proposed method can precisely identify pavement fractures while still maintaining the edge.