Researchers have developed a novel neural codec designed to improve the efficiency of transmitting imagery from Earth-observation satellites. This codec utilizes a clear-probability-weighted reconstruction loss, allowing it to reallocate coded bytes from cloudy areas to clear ground without needing a cloud map. The system encodes imagery into base and refinement layers, enabling prioritized transmission of clear content based on factors like deadline slack and remaining bytes. Evaluations show this approach requires significantly fewer bytes than existing methods for comparable clear-region quality and consumes less energy than traditional cloud detection methods. AI
IMPACT This research could lead to more efficient data transmission for satellite imagery, potentially enabling higher-resolution or more frequent data collection.
RANK_REASON Publication of a new research paper on arXiv detailing a novel technical approach. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
