IMAGE RENDERING AND TERRAIN GENERATION OF PLANETARY SURFACES USING SOURCE-AVAILABLE TOOLS

Simulating planetary surface imagery is critical for tasks such as mapping, optical navigation, and scientific exploration. It is also vital for training and validating autonomous surface nav- igation algorithms. Such applications necessitate high-fidelity synthetic images that authen- tically mirror the unique properties of planetary surfaces, including their reflectance, albedo, and topography. In this paper, we introduce rendering pipelines based on Blender Cycles and Unreal Engine 5, an open-source and source-available software tool, respectively. While Blender Cycles offers superior fidelity imagery via path tracing at the expense of compu- tational costs, Unreal Engine 5 provides real-time rendering capabilities with photorealistic results. Using these tools, we delve into the integration of user-defined reflectance models, such as the Hapke model, and explore procedural-terrain-generation techniques to create en- tirely synthetic celestial bodies. Our demonstrations include replicating an actual image from the Rosetta mission, rendering an image of the Moon, and producing procedurally-generated asteroids using Blender. Additionally, we use Unreal Engine 5 to create a procedural rubble- pile asteroid, comprising thousands of surface rocks and amounting to billions of triangular mesh elements, all rendered in real time. Despite some constraints, we conclude that these tools show promising potential as resources for training and testing autonomous navigation algorithms.