Mosaic Method for Bridge Surface Imaging Based on Image Correction and Path Optimization

Abstract

To address the fundamental issue of large-scale bridge surface defect detection by inspection robots, an innovative stitching method based on image correction and shooting path optimization is proposed. By presetting shooting angles, distances, cropping ratios, and camera parameters, reasonable shooting step lengths and paths are determined. Correction coefficients are calculated for each photo, and the photos undergo refined processing of correction followed by cropping according to the cropping ratio, eliminating distortions and redundant information during the shooting process. The cropped and corrected photos are then seamlessly stitched together in numerical order to generate complete and high-quality images. The stitching method in this study is not only suitable for complex scenes without prominent feature points but also significantly reduces computational resource consumption by simplifying calculation processes and optimizing algorithm design. A verification experiment was conducted on the Mozhai Wujiang Grand Bridge of the Chongqing-Hunan Expressway Duplication Project, successfully stitching together a complete image of the bridge’s main girder. Compared with traditional image stitching methods, this method significantly improves stitching efficiency and accuracy while ensuring the accuracy and reliability of the stitching results. This study provides an efficient and accurate image stitching technology for bridge surface defect detection, with broad application prospects and important practical value.