Ground Visibility in Satellite Optical Time Series Based on A Contrario Local Image Matching
Rafael Grompone von Gioi, Charles Hessel, Tristan Dagobert, Jean-Michel Morel, Carlo de Franchis
⚠ This is a preprint. It may change before it is accepted for publication.


Assessing ground visibility is a crucial step in automatic satellite image analysis. Some Earth observation satellites are provided with spectral bands specially designed for cloud detection. An alternative approach is to detect ground visibility by comparing locally the images in a temporal series: matching regions are necessarily cloud free. Indeed, the ground has persistent patterns that can be observed repetitively in the time series, while clouds change shape constantly. We describe here a ground visibility detection algorithm based on an a contrario local image matching method, coupled with an efficient greedy algorithm.