Model Overview
The Cloud Gap Fill model reconstructs missing reflectance values in Sentinel-2 L2A (10 m resolution) tiles affected by cloud cover. It uses contextual information from both the input tile (or tile of interest) and nearby Sentinel tiles captured on nearby dates to intelligently fill cloudy pixels by estimating the nature of the underlying area.
This capability is crucial for ensuring the reliability of satellite images used in industries like agriculture, forestry, environmental monitoring, etc.
The model currently supports filling in values for only four bands i.e Blue, Green, Red and Near Infra-Red with details below.
| Band Name | Description |
|---|---|
| Red (B04) | Red channel of the input image |
| Green (B03) | Green channel of the input image |
| Blue (B02) | Blue channel of the input image |
| Near Infra-Red (B08) | Near Infra-Red channel of the input image |
