MARTIAN: A Rendering Framework for Aerial Mars Imagery from HiRISE Orbital Data

2026-05-28 · Source: Computer Vision and Pattern Recognition · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Robotics & Autonomous Systems, Computer Vision & Pattern Recognition · Depth: Advanced, quick

Summary

MARTIAN is an open-source, Blender-based rendering framework designed to synthesize realistic aerial views of Martian terrain. It leverages real HiRISE orbital map products to generate imagery under controllable lighting conditions and at varying altitudes, complete with accurate pose annotations. This framework directly addresses the critical bottleneck of scarce large-scale, annotated aerial datasets needed for training and evaluating vision-based navigation pipelines on Mars. MARTIAN has been validated through its deployment in concurrent work on map-based localization systems for Ingenuity and future Mars rotorcraft. In these applications, deep image matchers trained with MARTIAN's synthetic data were successfully evaluated using real Mars imagery. The framework is publicly available at https://github.com/nasa-jpl/martian.

Key takeaway

For Computer Vision Engineers developing autonomous navigation systems for Mars missions, MARTIAN offers a critical solution to data scarcity. You should integrate this open-source Blender-based framework to generate diverse, annotated aerial datasets, significantly accelerating model training and evaluation. This enables robust vision-based pipelines for future Mars rotorcraft, validated by its successful deployment with Ingenuity's localization systems. Utilize MARTIAN to reduce reliance on limited real-world imagery and enhance system reliability.

Key insights

MARTIAN synthesizes realistic Mars aerial imagery with pose annotations, solving data scarcity for vision-based navigation.

Principles

• Leverage real orbital data for synthetic generation.
• Synthetic data can validate real-world systems.
• Controllable conditions enhance training diversity.

Method

MARTIAN uses Blender and HiRISE orbital map products to render Martian terrain views, adding controllable lighting, altitude variations, and accurate pose annotations.

In practice

• Train deep image matchers for Mars navigation.
• Evaluate rotorcraft localization systems.
• Generate diverse datasets for vision pipelines.

Topics

• Mars Exploration
• Aerial Navigation
• Synthetic Data Generation
• HiRISE Imagery
• Blender Framework
• Computer Vision
• Rotorcraft Localization

Code references

• nasa-jpl/martian

Best for: AI Scientist, Research Scientist, Robotics Engineer, Computer Vision Engineer, Machine Learning Engineer

Related on AIssential

• AlphaJet: Automated Conceptual Aircraft Synthesis via Disentangled Generative Priors and Topology-Preserving Evolutionary Search — AlphaJet automates conceptual aircraft design using generative AI and evolutionary search for real-time 3D synthesis.
• OmniLiDAR: A Unified Diffusion Framework for Multi-Domain 3D LiDAR Generation — OmniLiDAR unifies 3D LiDAR generation across diverse sensing domains using a text-conditioned diffusion framework.
• NASA’s Perseverance rover completes the first AI-planned drive on Mars — AI-planned drives on Mars demonstrate advanced autonomous navigation for deep space exploration.
• NeRF-based Spacecraft Reconstruction from Close-Range Monocular Imagery Under Illumination Variability and Pose Uncertainty — NeRFs can be extended with per-image embeddings and pose corrections to robustly reconstruct spacecraft under challenging conditions.
• From Seeing to Simulating: Generative High-Fidelity Simulation with Digital Cousins for Generalizable Robot Learning and Evaluation — WorldComposer generates diverse, high-fidelity simulation environments from real-world panoramas for robust robot learning.
• Aerial Vision-Language Navigation with a Unified Framework for Spatial, Temporal and Embodied Reasoning — A unified framework enables aerial VLN using only monocular RGB and language, outperforming prior RGB-only methods.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Computer Vision and Pattern Recognition.