Terrain maps and satellite paths improved with neural inverse graphics
OrTeN: Neural Inverse Graphics for Joint Terrain and Line-scan Camera Trajectory Reconstruction
Computer Vision and Pattern RecognitionGraphicsMachine Learning
Summary
Making detailed maps of planets' surfaces from satellite pictures is hard because the cameras can have errors and light conditions change. The authors created a new method called OrTeN that uses a type of AI to fix both the satellite’s path and the terrain map at the same time. OrTeN looks directly at satellite images and improves the map by making sure different views of the same spots match well. It also accounts for shadows and lighting changes over time. This helps get accurate 3D maps without needing extra starting information about the landscape.
What this means in practice
- •For planetary mission teams: Improve terrain maps and spacecraft positioning for planetary landings using only calibrated satellite images and initial camera poses.
- •For geological survey teams: Create accurate 3D terrain models from diverse satellite images without relying on prior depth data or rough terrain estimates.
Authors
Josef Biberstein, Sertac Karaman
Abstract
Digital terrain maps (DTMs) are used for landing site selection, terrain-relative navigation, and geological analysis in planetary exploration. Existing state-of-practice reconstruction techniques for creating DTMs from satellite imagery use multi-view stereo, requiring complex pipelines for dealing with camera error, changing illumination, and images with limited disparity. This work introduces Orbital Terrain Networks (OrTeN), a DTM construction method that uses neural inverse rendering to jointly optimize a textured height map and orbital camera corrections from line-scan satellite images. OrTeN operates directly on calibrated line-scan images with pixel values expressed as radiance factor (I/F) and corrects both broad trajectory pose inconsistencies and high-frequency pointing jitter jointly with surface reconstruction during training. Surface-projected patch normalized cross-correlation (NCC) enforces consistency between observations of the same terrain, refining an initial rough reconstruction. A semi-empirical reflectance model and shadow rendering are used to fit surface appearance from multi-date imagery with differing solar illumination. OrTeN requires only calibrated line-scan images and initial camera poses and does not use any structural priors including ranging/depth measurements or initial rough DTMs. OrTeN is evaluated on four planetary imagery datasets. Aligned terrain RMSEs are 2.35 m at Jezero Crater, 2.82 m at Gale Crater, 16.18 m at Olympus Mons, and 0.43 m at McGetchin Crater.