World’s First Fully Automatic Haze Removal and Image Clarification Method for Optical Satellite Imagery
AI_COOL makes satellite imagery clearer, more accurate, and more reliable.
AI_COOL Technology Overview
AI_COOL decomposes the electromagnetic transmission equation into three components — I(X): satellite observation, J(X): true surface radiance, and H(X): haze radiance — providing a clear physical foundation for haze removal and image clarification.
Using multiscale spatial analysis, fractal and other modern mathematical methods, AI_COOL automatically identifies the spatial structure of haze and separates the haze layer from the original image.
With AI-driven, fully automatic quality control, the system restores surface radiance without any parameter input, producing high-precision, haze-free images for quantitative remote sensing, land-cover analysis, and large-area monitoring.
AI_COOL supports rapid, batch-mode processing for optical sensors, including satellite, aerial, and drone platforms, offering a fast and reliable solution for complex and non-uniform haze conditions.
AI_COOL Key Features
- Physically based haze decomposition — Separates I(X) into J(X) and H(X) using the electromagnetic transmission equation.
- Fractal-guided haze structure detection — Identifies non-uniform haze using multiscale spatial and fractal analysis.
- Zero-parameter, fully automatic processing — No meteorological data, no atmospheric parameters, no manual tuning.
- AI-driven quality control — Automatically selects the optimal haze-removal solution.
- High-throughput batch processing — Maintains spectral fidelity across large-scale datasets.
- Multi-sensor compatibility — Works with all high-resolution optical satellite, aerial, and drone imagery.
- Haze distribution retrieval — Outputs haze intensity map H(X) for environmental applications.
- Robust under complex conditions — Effective for heavy, non-uniform, or spatially variable haze.
Why AI_COOL Is Different
Traditional atmospheric correction software requires many input parameters, expert tuning, and subjective adjustments. AI_COOL eliminates these limitations by directly estimating and removing the haze layer using modern mathematical tools and automated AI quality control — achieving true, high-accuracy, fully automatic haze removal.
AI_COOL vs ATCOR
ATCOR uses physical atmospheric modeling and requires parameter inputs such as visibility, aerosol type, and water vapor.
AI_COOL requires no inputs, uses spatial–fractal analysis to extract haze directly from the image, and includes automatic AI evaluation for clarity optimization.
Example Visual Results
Application Areas
- Land cover / land use monitoring
- Urban change detection
- Biodiversity monitoring
- Environmental management
- Crop analysis and precision agriculture
- Disaster response mapping