Title: Application of binocular image reconstruction method in the construction of 3D model of wooden arch corridor bridge structure

Authors: Hua Deng

Addresses: Department of Management, Chengyi College, Jimei University, Xiamen, 361021, China

Abstract: In order to better understand the structure of wooden-arch corridor bridges and preserve cultural buildings, the study proposes a three-dimensional model construction method based on binocular image reconstruction method. Facing the problem of poor image edge feature extraction, the study uses Laplace combined with the Sobel operator (Laplace-Sobel) for image edge feature extraction. The SIFT structural feature matching algorithm is improved to eliminate the mismatching points in stereo matching. The results show that as the active window size increases, the number of detected edge feature points gradually increases and the number of feature stereo matching pairs also increases, but the number of matching pairs is lower than the number of feature points; the number of features is searched for the most and the matching cost is the least when the neighbourhood range of pixels is 15 × 15. The results show that the improved RANSAC-SIFT matching algorithm has an average root-mean-square geometric error of 1.2482 and a high matching accuracy, with an average running time of 94 ms. Compared with the unimproved algorithm, the performance has been significantly improved and can be better applied in binocular image reconstruction, reducing the difficulty of constructing 3D models of wooden arch bridges.

Keywords: binocular image reconstruction method; three-dimensional model; wooden arch gallery bridge structure; matching algorithm.

DOI: 10.1504/IJCISTUDIES.2023.137811

International Journal of Computational Intelligence Studies, 2023 Vol.12 No.3/4, pp.223 - 237

Received: 19 Sep 2022
Accepted: 10 Mar 2023
Published online: 05 Apr 2024 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article