Skip to Main Content
 

Global Search Box

 
 
 
 

Files

File List


Shuang_PhD_Thesis_revised.pdf (15.2 MB)

ETD Abstract Container

Abstract Header

Scalable Scene Modeling from Perspective Imaging: Physics-based Appearance and Geometry Inference

Song, Shuang
http://orcid.org/0000-0002-0037-1499
http://rave.ohiolink.edu/etdc/view?acc_num=osu1704416076215084

Abstract Details

2024, Doctor of Philosophy, Ohio State University, Civil Engineering.
3D scene modeling techniques serve as the bedrocks in the geospatial engineering and computer science, which drives many applications ranging from automated driving, terrain mapping, navigation, virtual, augmented, mixed, and extended reality (for gaming and movie industry etc.). This dissertation presents a fraction of contributions that advances 3D scene modeling to its state of the art, in the aspects of both appearance and geometry modeling. In contrast to the prevailing deep learning methods, as a core contribution, this thesis aims to develop algorithms that follow first principles, where sophisticated physic-based models are introduced alongside with simpler learning and inference tasks. The outcomes of these algorithms yield processes that can consume much larger volume of data for highly accurate reconstructing 3D scenes at a scale without losing methodological generality, which are not possible by contemporary complex-model based deep learning methods. Specifically, the dissertation introduces three novel methodologies that address the challenges of inferring appearance and geometry through physics-based modeling. Firstly, we address the challenge of efficient mesh reconstruction from unstructured point clouds—especially common in large and complex scenes. The proposed solution employs a cutting-edge framework that synergizes a learned virtual view visibility with graph-cut based mesh generation. We introduce a unique three-step deep network that leverages depth completion for visibility prediction in virtual views, and an adaptive visibility weighting in the graph-cut based surface. This hybrid approach enables robust mesh reconstruction, overcoming the limitations of traditional methodologies and showing superior generalization capabilities across various scene types and sizes, including large indoor and outdoor environments. Secondly, we delve into the intricacies of combining multiple 3D mesh models, particularly those obtained through oblique photogrammetry, into a unified high-resolution site model. This novel methodology circumvents the complexity of traditional conflation by using a panoramic virtual camera field and Truncated Signed Distance Fields. The result is a seamless handling of full-3D mesh conflations, which is a daunting task in standard geoscience applications due to intricate topologies and manifold geometries. The developed technique substantially enhances the accuracy and integrity of resultant 3D site models, empowering geoscience and environmental monitoring efforts. Thirdly, the dissertation introduces a physics-based approach for the recovery of albedo from aerial photogrammetric images. This general albedo recovery method is grounded in a sophisticated inverse rendering framework that capitalizes on the specifics of photogrammetric collections—such as known solar position and estimable scene geometry—to recover albedo information accurately. The effectiveness of the approach is demonstrated through significant improvements not only in the realism of the rendered models for VR/AR applications but also in the primary photogrammetric processes. Such advancements have the potential to refine feature extraction, dense matching procedures, and the overall quality of synthetic environment creation. Overall, the research encapsulated in this dissertation marks a series of methodological triumphs in the processing of complex datasets. By navigating the confluence of deep learning, computational geometry, and photogrammetry, this work lays down a robust framework for future exploration and practical application in the rapidly evolving field of 3D scene reconstruction. The outcomes of these studies are evidenced through rigorous experiments and comparisons with existing state-of-the-art methods, demonstrating the efficacy and scalability of the proposed approaches.
Rongjun Qin (Advisor)
Alper Yilmaz (Committee Member)
Charles Toth (Committee Member)
112 p.
Civil Engineering
3D scene modeling; Physics-based modeling; Mesh reconstruction; Oblique photogrammetry; Albedo recovery

Recommended Citations

Citations

  • Song, S. (2024). Scalable Scene Modeling from Perspective Imaging: Physics-based Appearance and Geometry Inference [Doctoral dissertation, Ohio State University]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=osu1704416076215084

    APA Style (7th edition)

  • Song, Shuang. Scalable Scene Modeling from Perspective Imaging: Physics-based Appearance and Geometry Inference . 2024. Ohio State University, Doctoral dissertation. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=osu1704416076215084.

    MLA Style (8th edition)

  • Song, Shuang. "Scalable Scene Modeling from Perspective Imaging: Physics-based Appearance and Geometry Inference ." Doctoral dissertation, Ohio State University, 2024. http://rave.ohiolink.edu/etdc/view?acc_num=osu1704416076215084

    Chicago Manual of Style (17th edition)

osu1704416076215084
31
© 2024, all rights reserved.
This open access ETD is published by The Ohio State University and OhioLINK.