Large Scale Photometric Bundle Adjustment
This work addresses the challenge of large-scale photometric bundle adjustment for 3D reconstruction from diverse internet photos, offering incremental improvements over existing methods.
The paper tackles the problem of offline 3D reconstruction from internet images by jointly optimizing dense geometry and camera parameters using a photometric cost invariant to lighting changes, resulting in improved metric reconstruction accuracy over feature-based methods on the Tanks & Temples benchmark.
Direct methods have shown promise on visual odometry and SLAM, leading to greater accuracy and robustness over feature-based methods. However, offline 3-d reconstruction from internet images has not yet benefited from a joint, photometric optimization over dense geometry and camera parameters. Issues such as the lack of brightness constancy, and the sheer volume of data, make this a more challenging task. This work presents a framework for jointly optimizing millions of scene points and hundreds of camera poses and intrinsics, using a photometric cost that is invariant to local lighting changes. The improvement in metric reconstruction accuracy that it confers over feature-based bundle adjustment is demonstrated on the large-scale Tanks & Temples benchmark. We further demonstrate qualitative reconstruction improvements on an internet photo collection, with challenging diversity in lighting and camera intrinsics.