ExploreGS: a vision-based low overhead framework for 3D scene reconstruction
This addresses the problem of costly and resource-intensive 3D reconstruction for drone applications, offering a more accessible solution, though it appears incremental as it builds on existing vision and 3DGS techniques.
The paper tackles 3D scene reconstruction for drones by proposing ExploreGS, a vision-based framework that replaces lidar with RGB images, achieving high-quality reconstruction at lower cost and enabling real-time on-board processing with 3D Gaussian Splatting, while maintaining quality comparable to state-of-the-art methods.
This paper proposes a low-overhead, vision-based 3D scene reconstruction framework for drones, named ExploreGS. By using RGB images, ExploreGS replaces traditional lidar-based point cloud acquisition process with a vision model, achieving a high-quality reconstruction at a lower cost. The framework integrates scene exploration and model reconstruction, and leverags a Bag-of-Words(BoW) model to enable real-time processing capabilities, therefore, the 3D Gaussian Splatting (3DGS) training can be executed on-board. Comprehensive experiments in both simulation and real-world environments demonstrate the efficiency and applicability of the ExploreGS framework on resource-constrained devices, while maintaining reconstruction quality comparable to state-of-the-art methods.