ABot-Earth 0.5: Generative 3D Earth ModelAuthors: ABot Team (AutoDrive/Amap) |Year: 2026 |arXiv: 2606.09967二、研究背景与动机传统大规模 3D 重建依赖密集倾斜摄影测量和 LiDAR数据采集成本极高、处理延迟长、计算门槛高制约了实时或按需的全球建模。生成式 3D 建模是绕过这些物理限制的有效路径已在单体对象尺度TRELLIS、Hunyuan3D 等取得巨大成功。然而将对象级方法扩展到无边界室外场景面临三大瓶颈表示鸿沟真实环境需要非流形拓扑、规模与交互性地球级需无缝 LOD、空间一致性大范围拼接产生伪影。图1ABot-Earth 从卫星图像生成的全球各地城市与自然地形 3D 场景示例。四、实验4.1 生成保真度学术对比与 CityDreamer、GaussianCity、EarthCrafter 对比使用真实世界 3DGS 重建的渲染作为 ground-truth 分布方法FID ↓KID ↓CityDreamer~200—GaussianCity~100—EarthCrafter69.5—ABot-Earth 0.516.1最优FID 16.1 相比前最优 69.5 提升显著且评测基准为真实世界复杂 3DGS 重建渲染难度远高于合成数据集。4.2 系统级对比商业解决方案与 Google Earth摄影测量重建和 Marble闭源程序化生成对比评估四个维度图5系统对比雷达图几何精度/纹理保真度/整体美感及覆盖范围对比。图6爱尔兰地区对比——ABot-Earth 生成完整 3D 场景Google Earth 因缺乏扫描数据仅显示 2D 图像。关键差异空间覆盖Google Earth 3D 资产仅覆盖少数发达国家 CBDABot-Earth 可生成任意坐标效率ABot-Earth 10 分钟/km²Google Earth 更新周期为数月至数年美学评分ABot-Earth 高于 Google Earth整体光照和色彩和谐性更好开放性原生 3DGS 输出支持任意视角渲染和下游集成4.3 地标增强混合生成-重建用 COLMAP 对地标埃菲尔铁塔、罗马斗兽场、美国国会大厦、凯旋门重建高精度 3DGS地理配准后合成到生成环境中图7重建地标合成到生成环境的结果。保留精细建筑细节与周围生成环境自然融合。4.4 下游应用图8基于 ABot-Earth 生成环境的无人机UAV导航与避障仿真演示。图9奥克兰城市场景ABot-Earth 生成60°俯仰角。图10奥克兰城市场景Google Earth同视角对比。六、总结ABot-Earth 0.5 代表了生成式 3D 建模从对象级到地球级的关键跨越。其核心价值在于以卫星图像为条件无需大量多视角数据在 10 分钟/km² 的效率下生成 FID 16.1 的高保真 3DGS 场景支持行星级交互式探索。这为目前 Google Earth 等商业方案无法覆盖的大量区域发展中国家城市、非 CBD 区域、自然地形提供了可扩展的低成本替代方案有望推动全球地理空间技术的普惠化。报告生成时间2026-06-11 | 论文来源arXiv:2606.09967原文摘要: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.PDF链接:https://arxiv.org/pdf/2606.09967v1部分平台可能图片显示异常请以我的博客内容为准