A Cloud-Contamination Removal Algorithm by Combining Image Segmentation and Level-set-based Approaches for Remote Sensing Images

摘要

This paper proposes a cloud removal algorithm for satellite images that are obscured by clouds. Our concept is to separate the land cover map into the segments. After that, we localize and indicate the degree of cloud thickness map by using a level set function. As the continuity of the level set function, the gradient approach is deployed by using the calculus of variation. Our example shows that our model can mitigate the effect clouds in images.