AeroLLE: Constrained Pseudo-Supervision for Nighttime Aerial Image Enhancement with the AeroNight-1.5K Benchmark
For aerial imaging applications, this provides a practical strategy for learning from generated appearance guidance when registered references are unavailable, but the gains are incremental over existing enhancement methods.
The paper tackles nighttime aerial image enhancement, where registered normal-light references are unavailable. They introduce the AeroNight-1.5K benchmark with 1,500 real images and propose AeroLLE, a two-stage framework with a base enhancer and spatially adaptive calibration, showing improved agreement with screened appearance targets and more balanced exposure/color correction.
Nighttime aerial image enhancement is challenged by spatially nonuniform exposure, mixed illumination, and weak structural evidence, while registered normal-light targets are difficult to capture from moving platforms. Generated normal-light images provide practical appearance guidance but may alter geometry or texture. We introduce \aeronight{}, comprising 1,500 real nighttime aerial RGB images: 1,300 inputs are associated with manually screened pseudo-references, and 200 inputs support unpaired evaluation. We propose AeroLLE, a two-stage framework that first recovers visibility with an HVI Base Enhancer and then performs Spatially Adaptive Exposure--Color Calibration (SAECC). After the Base Enhancer is selected and frozen, SAECC predicts bounded, low-resolution RGB gain and bias fields, restricting the magnitude and spatial variation of the second-stage correction. Experiments under complementary pseudo-paired and unpaired protocols demonstrate improved agreement with screened appearance targets, together with more balanced exposure and color correction across diverse nighttime aerial scenes. These results support constrained, stage-specific calibration as a practical strategy for learning from generated appearance guidance when registered aerial references are unavailable.