A binocular reconstruction method fused with Laplacian image information for pavement texture evaluation

Abstract

3D reconstruction technology based on image processing has the advantages of fast speed and high resolution to reconstruct pavement texture morphology for detecting the skid resistance. However, the small color difference and unobvious feature points of asphalt pavement lead to low accuracy in texture morphology reconstruction. A binocular technology fused with Laplacian image information was proposed to improve the precision of texture reconstruction as follow: the first mode without region segmentation, the second mode with laser region segmentation and the third mode with the region segmentation used template. The results show that the proposed algorithm improves the binocular reconstruction accuracy of pavement texture compared with the traditional SGBM algorithm. The proposed method and the laser scanner are compared through areal texture parameters. The results of the proposed method are relatively close to those of laser scanners.