Tile and Slide : A New Framework for Scaling NeRF from Local to Global 3D Earth Observation

TL;DR

This paper presents Snake-NeRF, a scalable out-of-core framework for large-scale 3D reconstruction from satellite imagery, enabling efficient processing on a single GPU without quality loss.

Contribution

Snake-NeRF introduces a novel tiling and sampling strategy that allows large satellite scenes to be reconstructed efficiently on a single device.

Findings

Effective large-scale satellite image processing with linear time complexity.

No compromise in reconstruction quality for large scenes.

Operates on a single GPU without loading all images and networks simultaneously.

Abstract

Neural Radiance Fields (NeRF) have recently emerged as a paradigm for 3D reconstruction from multiview satellite imagery. However, state-of-the-art NeRF methods are typically constrained to small scenes due to the memory footprint during training, which we study in this paper. Previous work on large-scale NeRFs palliate this by dividing the scene into NeRFs. This paper introduces Snake-NeRF, a framework that scales to large scenes. Our out-of-core method eliminates the need to load all images and networks simultaneously, and operates on a single device. We achieve this by dividing the region of interest into NeRFs that 3D tile without overlap. Importantly, we crop the images with overlap to ensure each NeRFs is trained with all the necessary pixels. We introduce a novel $2\times 2$ 3D tile progression strategy and segmented sampler, which together prevent 3D reconstruction errors along the tile edges. Our experiments conclude that large satellite images can effectively be processed with linear time complexity, on a single GPU, and without compromise in quality.