Sat2City v2: Native 3D City Asset Generation from a Single Satellite Image

2026-06-23

AI Summary

This work proposes a novel method for generating reusable, textured explicit 3D urban assets from a single satellite image, targeting applications in digital twins and urban simulation. Leveraging a pre-trained native 3D latent diffusion model, the approach introduces satellite-conditioned geometric flow fine-tuning and a cascaded geometry-to-appearance generation mechanism, enabled by weakly aligned satellite-mesh data pairs. This paradigm shift moves from rendering-oriented proxies to explicit mesh assets suitable for real-world deployment. The study presents the first large-scale real-world dataset comprising 16,241 satellite–3D mesh pairs and demonstrates superior performance over existing methods in both geometric fidelity and visual appearance, significantly enhancing the quality and practical utility of the generated 3D urban assets.

Abstract

Generating explicit 3D city assets from a single satellite image is important for digital twins, urban simulation, and geospatial intelligence. Unlike satellite-to-street-view synthesis, the task requires a reusable textured mesh with plausible geometry and controllable appearance rather than a 3D proxy optimized only for rendering a small set of images or videos. The ICCV Sat2City framework made a first step by conditioning cascaded sparse-voxel latent diffusion on satellite-derived height maps, but its appearance was random, its training data were synthetic, and its task-specific VAE did not scale well to noisy real-world reconstructions. We present Sat2City v2, a journal extension that adapts a pretrained native structured-latent 3D foundation model to weakly aligned satellite images and textured meshes. We build a real-world dataset with 16,241 satellite-mesh pairs across 24 regions in 9 cities. Instead of learning a 3D representation from noisy city meshes, Sat2City v2 encodes each mesh into a pretrained native 3D latent space, fine-tunes a satellite-conditioned geometry flow, and uses the decoded shape to anchor satellite-conditioned texturing. This retains Sat2City's geometry-to-appearance cascade while enabling appearance-controllable generation from the satellite input. Experiments on metric-scale DSM reconstruction and generative city-asset benchmarks for geometry and appearance show that Sat2City v2 achieves the best overall performance among evaluated baselines. Overall, Sat2City v2 advances satellite-to-city generation from rendering-oriented 3D proxies to explicit textured mesh assets, supported by, to the best of our knowledge, the first documented satellite-mesh paired dataset collected from matched geographic crops for this asset-level task. Project page: https://ai4city-hkust.github.io/Sat2City-v2/

Problem

• 3D city asset generation
• satellite image
• textured mesh
• geometry and appearance control
• digital twins

Innovation

• 3D city asset generation
• satellite-to-mesh
• structured latent diffusion
• geometry-aware texturing
• real-world satellite-mesh dataset