Session:
Session 2: Quality Control, Product Validation, Best Practices 1
Title:
Lemu Nge preprocessing and initial in-flight calibration performances during the commissioning
Abstract:
In this presentation, the overall Lemu Nge preprocessing and initial in-flight calibration activities and results will be presented. We gratefully acknowledge the Lemu team for their support and for providing access to the satellite datasets.
Launched in 2024 and led from the Chilean Patagonia, Lemu Nge is the world’s first satellite mission dedicated exclusively to biodiversity monitoring. The 6U nanosatellite (30 × 20 × 10 cm) carries a Simera Sense 32-band hyperspectral sensor covering the 420–900 nm spectral range, with a ground sampling distance of ~4.75 m. Operating in low Earth orbit at ~600 km altitude, the satellite revisits specific areas every 3–7 days, enabling systematic monitoring of diverse ecosystems including forests, wetlands, and coastal zones, as well as detecting ecosystem change and habitat condition. The generated datasets feed into Atlas, Lemu’s Nature Intelligence Platform, where they enhance ecological indicators such as LemuRank (spatial conservation prioritization) and Biorarity (rarity-weighted species richness). These indicators are already in operational use by diverse organizations in Chile.
The scientific utility of the mission relies on rigorous data pre-processing and in-flight calibration, provided as a service by VITO in partnership with Lemu. Following launch and the successful completion of the Launch and Early Orbit Phase (LEOP), NanoAvionics commissioned the platform and subsystems to confirm nominal functionality. Subsequently, calibration acquisitions over designated reference sites enabled radiometric and geometric in-flight assessments. VITO developed a continuous and autonomous pre-processing chain, delivered as software to Lemu, which systematically transforms raw imagery into standardised datasets. In parallel, the VITO CalibrEO service was used to monitor and optimize radiometric and geometric performance.
Through in-flight calibration and validation, the absolute geolocation accuracy was improved from approximately 1 km to sub-pixel level by estimating corrections to the exterior orientation (boresight angles) and interior orientation parameters (CCD line-of-sight vectors). Systematic spectral band alignment further ensured precise band co-registration (below 1/3 of pixel).
Radiometric processing started from pre-flight calibration and characterisation data. The presence of both vertical and horizontal striping necessitated the implementation of a destriping algorithm, systematically applied to improve image quality. Initial in-flight radiometric calibration against Sentinel-2 and desert PICS revealed a consistent underestimation of TOA reflectance, most pronounced in the blue region. These results highlight the importance of vicarious radiometric calibration to ensure absolute radiometric accuracy.
These calibration and validation activities will guarantee that Lemu Nge provides high-quality, scientifically robust hyperspectral data products suitable for ecological research, conservation planning, and operational monitoring.
Sindy Sterckx