Abstract
This paper presents an accurate and robust dense 3D reconstruction system for detail preserving surface modeling of large-scale scenes from multi-view images, which we named DP-MVS. Our system performs high-quality large-scale dense reconstruction, which preserves geometric details for thin structures, especially for linear objects. Our framework begins with a sparse reconstruction carried out by an incremental Structure-from-Motion. Based on the reconstructed sparse map, a novel detail preserving PatchMatch approach is applied for depth estimation of each image view. The estimated depth maps of multiple views are then fused to a dense point cloud in a memory-efficient way, followed by a detail-aware surface meshing method to extract the final surface mesh of the captured scene. Experiments on ETH3D benchmark show that the proposed method outperforms other state-of-the-art methods on F1-score, with the running time more than 4 times faster. More experiments on large-scale photo collections demonstrate the effectiveness of the proposed framework for large-scale scene reconstruction in terms of accuracy, completeness, memory saving, and time efficiency.
Highlights
Multi-view stereo (MVS) reconstruction of large-scale scenes is a research topic of vital importance in computer vision and photogrammetry
Inspired by [26], we propose a detail preserving PatchMatch method based on the diffusion-like propagation scheme, which ensures both high accuracy and completeness of the estimated depth map, especially for accurate reconstruction of detailed structures
From the experimental results we can see that our DP-MVS approach performs better than the other methods in the generated 3D models, especially in those regions which contain rough surface structures and thin structures, which validates the effectiveness of our DPMVS method
Summary
Multi-view stereo (MVS) reconstruction of large-scale scenes is a research topic of vital importance in computer vision and photogrammetry. With the development of smart cities and digital twin, 3D reconstruction of large-scale scenes has attracted more attentions due to its usefulness in providing digitalized content for various applications such as urban visualization, 3D navigation, geographic mapping, and model vectorization. These applications usually require reconstruction of high-quality dense surface models. 3D visualization and navigation demand realistically textured 3D surface models with complete structures and few artifacts, while geographic mapping and model vectorization depends on highly accurate dense point clouds or models with geometric details as reliable 3D priors, which are great challenges to multi-view reconstruction.
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
