Researchers have introduced AeroLLE, a two-stage framework designed to enhance nighttime aerial images. This method addresses challenges such as uneven lighting and the difficulty of capturing aligned normal-light references from moving platforms. The framework first improves visibility using a base enhancer and then applies Spatially Adaptive Exposure--Color Calibration (SAECC) to refine exposure and color balance. Experiments show that AeroLLE effectively learns from generated appearance guidance, even when registered references are unavailable, by employing constrained, stage-specific calibration. AI
IMPACT This research offers a new approach to image enhancement for challenging nighttime aerial scenes, potentially improving applications in surveillance and mapping.
RANK_REASON This is a research paper detailing a new method for image enhancement. [lever_c_demoted from research: ic=1 ai=1.0]
- AeroLLE
- AeroNight-1.5K
- arXiv
- Hugging Face
- HVI Base Enhancer
- SAECC
- Spatially Adaptive Exposure--Color Calibration
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →