ABot-Earth 0.5: Generative 3D Earth Model

Popular Boards

Preprint 1 Mention

Ming Qian•2026

Tianjian OuyangMingchao Sun

Few Citations

0 citations · Environmental Engineering

Open Access

TLDR

This paper introduces a new tool that quickly turns satellite images into realistic 3D maps, making it easier and cheaper for anyone to create and explore digital versions of real places.

Summary

AI Generated

This content was automatically synthesized by Credo's AI models directly from the original source text.

AI summaries can make mistakes — double-check important details against the original source.

1 Study Aim

The main goal of this paper is to introduce ABot-Earth 0.5, a new system that can generate large, realistic 3D environments using only satellite images. The authors aim to make high-quality 3D reconstructions more accessible and affordable for a wide range of users. Simply put: The study wants to make it easy and cheap for anyone to create lifelike 3D maps from satellite pictures.

2 Study Design

The researchers developed a generative model based on 3D Gaussian Splatting (3DGS), which is a way to represent 3D shapes and textures using overlapping blobs of color and light. They trained this model on a large and varied set of real-world city reconstructions. The system only needs satellite images as input and can create new 3D scenes in less than 10 minutes for each square kilometer. The output includes different levels of detail, so users can view the 3D maps interactively and in real time on web-based platforms. Simply put: The team built and tested a computer program that learns from real 3D city models and then makes new 3D maps from satellite photos very quickly.

3 Findings

The study demonstrates that ABot-Earth 0.5 can produce highly realistic 3D environments from satellite images alone, at a much lower cost and faster speed than traditional methods. The generated 3D scenes are detailed enough for real-time use in web browsers and can support advanced applications like training artificial intelligence for drone navigation. The authors suggest that their approach removes many technical and financial barriers, making large-scale 3D mapping more practical for many users. They recommend using this tool to help bridge the gap between digital simulations and real-world environments. Simply put: The results show that this tool makes it much easier and cheaper to build and use detailed 3D maps from satellite images.

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.

Primary Link Archive DOI Openalex Full text (OA)