Preparing for DAVINCI Through Multi-Scale Drone Reconnaissance of Icelandic Lava Shields

Introduction: In the early 2030s, NASA’s DAVINCI mission will capture unprecedented near-infrared descent imagery of the Alpha Regio tesserae on Venus. The Venus Descent Imager (VenDI), built by MSSS, is designed to acquire high-resolution optical data from ~21 km down to just 200–400 meters before touchdown [1]. This dataset will allow scientists to rigorously assess geological processes, chronological relationships, and end-member rock compositions. However, because the planetary science community currently lacks any high-resolution optical imagery of Venus, there is no existing analytical framework for interpreting these sub-cloud observations over complex volcanic landscapes.

To bridge this critical gap and develop operational strategies for VenDI, we executed a comprehensive uncrewed aerial systems (UAS) campaign at the Sandfellshaed lava shield and other Reykjanes Peninsula sites in Iceland. By utilizing Structure-from-Motion (SfM) techniques to create high-resolution Digital Elevation Models (DEMs) and orthomosaics at centimeter to sub-centimeter scales [2, 3], these analog sites act as critical "training grounds." They provide the necessary ground-truth context to establish the spatial resolution limits required to recognize primary geologic features during DAVINCI's descent, ultimately maximizing the mission's scientific return.

Field Campaign Methodology: Over four days of active fieldwork, our team used UAS platforms to gather georeferenced images across approximately 1 km² of the Sandfellshaed shield and fresh Svartsengi lavas. Flight patterns were specifically engineered to simulate DAVINCI’s operational profiles: (1) broad terrain mapping at a 4 cm/pixel ground sample distance for regional DEM creation; (2) simulated descent trajectories from 120 m above ground level (the maximum permissible altitude) to mimic DAVINCI's approach angles; (3) Super-Low Altitude Mode (SLAM) traverses at 0.5–1 cm/pixel to capture sub-centimeter details, similar to the Curiosity rover's MARDI camera; and (4) highly targeted imaging of specific volcanic structures, such as radial lava flows, structural fractures, and the summit crater. We processed more than 800 overlapping image frames using both Agisoft Metashape and custom NASA SfM/BPS pipelines tailored for planetary science. This yielded multi-scale DEMs spanning from a 5 cm grid scale for regional maps down to sub-centimeter local patches derived from SLAM flights.

Venus Analog Justification: The monogenetic Sandfellshaed lava shield serves as a premier terrestrial counterpart to the valley-filling volcanics and structurally complex terrains anticipated within Alpha Regio at sub-meter scales. This basaltic shield displays multi-scale roughness, morphological features, and surface textures that closely match theoretical models and existing Magellan SAR observations of Venus. Featuring a mix of structural fissures (gja), rough volcanic flow textures, and smooth lava expanses, this location provided the ideal environment to determine the feature recognition thresholds and scale dependencies necessary for VenDI's mission objectives.

Initial Results and Data Products: The Icelandic deployment yielded crucial benchmark datasets, including flight telemetry, simulated descent sequences, and multi-scale topography. These products are vital for interpreting future Venusian surface data that cannot be resolved from orbit. Key deliverables include highly accurate DEMs covering the entire shield complex at a ~6 cm grid scale (Figure 1), ultra-high-resolution orthomosaics at 3.2 cm/pixel, and detailed surface property evaluations comparing satellite data with ground-truth measurements. We successfully established the precise spatial scales (centimeters to meters) required to discriminate various volcanic features. Furthermore, we demonstrated that fusing derived topography with orthorectified imagery will enable the identification of primary geological processes, directly fulfilling DAVINCI’s Level-1 science requirements for surface analysis.

Implications for Venus Exploration: The analog studies conducted in SW Iceland define the operational detection limits needed to identify specific volcanic surface properties during DAVINCI's descent. By bridging the massive resolution gap between orbital radar (such as the tens-of-meters resolution expected from ESA’s EnVision) and localized lander observations, this multi-scale approach generates essential "training data" for scientific interpretation and automated feature recognition algorithms (e.g., AI/LLM tools). Ultimately, these resulting datasets and methodologies will optimize VenDI's descent imaging strategy while establishing analytical protocols applicable to future comparative planetary studies across both Venus and Mars.