We at Little Place Labs (LPL) are thrilled to announce the successful in-orbit demonstration of Orbitfy, our innovative product suite, as part of the first edition of the Cloud Computing Call for Ideas by ESA.
🌏 Our Orbitfy-Edge application made significant strides in processing satellite imagery on board D-Orbit's ION SCV004. It performed land use land cover (LULC) image segmentation, utilizing an advanced machine learning (ML) model we developed here at LPL. The app was specifically focused on identifying urban built-up areas, and it only downlinked the ML-segmented mask and a comprehensive report for images with an urban built-up area exceeding a set threshold. This approach drastically reduced the necessity to download extraneous imagery and preserved valuable downlink bandwidth.
📡 The results of this experiment were remarkable, achieving an astounding 98% reduction in the need to downlink imagery. The required analytics were minimized to just a few kilobytes. Even more impressively, this approach allowed for decision-making at the point of data collection, eliminating any latency.
👏 A hearty congratulations to our entire Little Place Labs team on this significant achievement!
We would like to express our gratitude towards D-Orbit and Unibap AB for organizing this initiative and to the European Space Agency - ESA - ESA Φ-lab for their continued support.
Check out the LinkedIn post by D-Orbit here.
To know more about LPL's solutions, please checkout https://www.littleplace.com/platform
留言