Batch-based Model Registration for Fast 3D Sherd Reconstruction
ICCV 2023
Abstract
3D reconstruction techniques have widely been used for digital documentation of archaeological fragments. However, efficient digital capture of fragments remains as a challenge. In this work, we aim to develop a portable, high-throughput, and accurate reconstruction system for efficient digitization of fragments excavated in archaeological sites. To realize high-throughput digitization of large numbers of objects, an effective strategy is to perform scanning and reconstruction in batches. However, effective batch-based scanning and reconstruction face two key challenges: 1) how to correlate partial scans of the same object from multiple batch scans, and 2) how to register and reconstruct complete models from partial scans that exhibit only small overlaps. To tackle these two challenges, we develop a new batch-based matching algorithm that pairs the front and back sides of the fragments, and a new Bilateral Boundary ICP algorithm that can register partial scans sharing very narrow overlapping regions. Extensive validation in labs and testing in excavation sites demonstrate that these designs enable efficient batch-based scanning for fragments. We show that such a batch-based scanning and reconstruction pipeline can have immediate applications on digitizing sherds in archaeological excavations.
1. Fast imaging process
To capture a group of fragments in a batch mode, we place them flat on the turntable, and first take a set of pictures to capture their exposed (front) sides. Then the fragments are flipped manually to photograph their back sides.
2. Reconstruction pipeline
Our reconstruction pipeline mainly has three steps: (1) Reconstruct partial 3D models from images; (2) Match front- and back-sides of fragments from scan batches; (3) Register the two sides of a sherd into a complete 3D model.
3. Dataset and reconstructed models
We built a dataset containing 123 fragments of different geometries, sizes, and textures. Then we validate the proposed system with a throughput of more than 700 fragments per day on this dataset.
4. Deployment on the site
The FIRES system was deployed for use at the excavation site of in Armenia in the summer of 2022 for two and half months.
Citation
@article{wang2023sherd,
title={Batch-based Model Registration for Fast 3D Sherd Reconstruction},
author={Wang, Jiepeng and Zhang, Congyi and Wang, Peng and Li, Xin and Cobb, Peter J. and Theobalt, Christian and Wang, Wenping},
booktitle = {International Conference on Computer Vision (ICCV)},
year = {2023}
}