ABot-Earth 0.5: Generative 3D Earth Model

2026-06-08

AI Summary

This work addresses the high cost, low efficiency, and limited real-time interactivity of large-scale, high-fidelity 3D Earth reconstruction by proposing a generative 3D Gaussian Splatting (3DGS) modeling approach driven by satellite imagery. It extends 3DGS to a global scale for the first time, integrating conditional generation with a hierarchical level-of-detail (LOD) structure to enable efficient, low-cost, and photorealistic synthesis of seamless large-area 3D scenes. The system can generate realistic 3D environments at a rate of one square kilometer within ten minutes and supports real-time web-based rendering as well as embodied AI applications such as closed-loop drone navigation. This significantly narrows the domain gap between simulation and reality while substantially lowering both the technical and economic barriers to large-scale 3D reconstruction.

Abstract

We present ABot-Earth 0.5, a generative 3D framework designed to synthesize vast, seamless 3D environments from ubiquitous, geospatially referenced satellite imagery. To achieve this, we propose a novel generative model formulated directly with the 3D Gaussian Splatting (3DGS) representation. The model is trained on a diverse corpus of existing real-world urban reconstructions, learning to generate realistic geometry and textures. At inference, it synthesizes novel 3D scenes conditioned solely on satellite imagery at a scalable rate of under 10 minutes per square kilometer, while demonstrating exceptional realism. The framework is designed for accessibility, with integrated hierarchical level-of-detail (LOD) structures that permit real-time, interactive visualization on web-based map engines. This high-fidelity simulation sandbox effectively mitigates the sim-to-real domain gap, enabling critical downstream Embodied AI applications like closed-loop UAV navigation. By providing an ultra-low-cost and high-efficiency solution, ABot-Earth 0.5 significantly lowers the technical and financial barriers to large-scale 3D reconstruction and empowers the future of global digital earth visualization.

Problem

• Generative 3D Earth
• Satellite Imagery
• 3D Reconstruction
• Sim-to-Real Gap
• Embodied AI

Innovation

• 3D Gaussian Splatting
• generative 3D modeling
• satellite imagery conditioning
• level-of-detail (LOD)
• Embodied AI