Construction of Digital Terrain Maps from Multi-view Satellite Imagery using Neural Volume Rendering

2025-08-02

AI Summary

Existing satellite multi-view stereo (MVS) methods rely on labor-intensive manual preprocessing and precise camera calibration, limiting their scalability and accuracy for large-area, high-fidelity terrain modeling in planetary exploration. To address this, we propose Neural Terrain Mapping (NTM), the first end-to-end digital terrain model (DTM) reconstruction method that requires only multi-view satellite imagery and coarse observation positions—without depth supervision, geometric priors, or accurate intrinsic/extrinsic camera parameters. NTM integrates differentiable projection with implicit volumetric representation within a neural volume rendering framework, enabling fully unsupervised training. Evaluated on both synthetic and real-world datasets, NTM reconstructs high-accuracy DTMs at near-original image resolution (sub-meter) over ~100 km² areas. Its performance matches that of state-of-the-art supervised MVS pipelines while significantly improving automation, robustness, and generalizability across complex planetary environments—including Earth and Mars.

Technology Category

• Computer Vision: Remote Sensing / Geospatial AI
• Machine Learning: Large Multimodal Models (LMMs)
• Intelligent Robots: Localization, Mapping, and Navigation

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Semantics and Knowledge: Methods, algorithms and applications for the development of semantic models, knowledge graphs and other forms of structured data models with machine-interpretable semantics
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Digital terrain maps (DTMs) are an important part of planetary exploration, enabling operations such as terrain relative navigation during entry, descent, and landing for spacecraft and aiding in navigation on the ground. As robotic exploration missions become more ambitious, the need for high quality DTMs will only increase. However, producing DTMs via multi-view stereo pipelines for satellite imagery, the current state-of-the-art, can be cumbersome and require significant manual image preprocessing to produce satisfactory results. In this work, we seek to address these shortcomings by adapting neural volume rendering techniques to learn textured digital terrain maps directly from satellite imagery. Our method, neural terrain maps (NTM), only requires the locus for each image pixel and does not rely on depth or any other structural priors. We demonstrate our method on both synthetic and real satellite data from Earth and Mars encompassing scenes on the order of $100 extrm{km}^2$. We evaluate the accuracy of our output terrain maps by comparing with existing high-quality DTMs produced using traditional multi-view stereo pipelines. Our method shows promising results, with the precision of terrain prediction almost equal to the resolution of the satellite images even in the presence of imperfect camera intrinsics and extrinsics.

Problem

• Improving digital terrain map construction from satellite imagery
• Reducing manual preprocessing in multi-view stereo pipelines
• Enhancing terrain prediction accuracy with neural rendering

Innovation

• Neural volume rendering for terrain maps
• Direct learning from satellite imagery
• No reliance on depth or structural priors