3DGS-ReLoc: 3D Gaussian Splatting for Map Representation and Visual ReLocalization
This addresses visual relocalization for robotics or autonomous systems, but appears incremental as it combines existing techniques like 3D Gaussian Splatting with standard optimization methods.
The paper tackles 3D mapping and visual relocalization by using LiDAR and camera data with 3D Gaussian Splatting to create detailed maps, achieving efficient localization via normalized cross-correlation and pose refinement on the KITTI360 dataset.
This paper presents a novel system designed for 3D mapping and visual relocalization using 3D Gaussian Splatting. Our proposed method uses LiDAR and camera data to create accurate and visually plausible representations of the environment. By leveraging LiDAR data to initiate the training of the 3D Gaussian Splatting map, our system constructs maps that are both detailed and geometrically accurate. To mitigate excessive GPU memory usage and facilitate rapid spatial queries, we employ a combination of a 2D voxel map and a KD-tree. This preparation makes our method well-suited for visual localization tasks, enabling efficient identification of correspondences between the query image and the rendered image from the Gaussian Splatting map via normalized cross-correlation (NCC). Additionally, we refine the camera pose of the query image using feature-based matching and the Perspective-n-Point (PnP) technique. The effectiveness, adaptability, and precision of our system are demonstrated through extensive evaluation on the KITTI360 dataset.