Satellite Imagery Clouds Cover Removal in ArcGIS Desktop

Cloud cover is a long-time problem for optical imagery analysis in remote sensing. It is obscuring important earth surface object that became the main target on image analysis. There are numerous methods and efforts have been developed to fix the cloud cover problem. The solutions are coming from simple approaches like masking and replacing using a raster calculator or map algebra, to more complex ways like multitemporal pixels blending, fast Fourier transform-based filtering, or even machine learning methods. 

The correction results also vary. Sometimes certain method could give a nearly perfect cloudless image, but sometimes it doesn't. One simple way to remove cloud in optical imagery is using the masking and replacing method. This method needs at least two imagery covering the same area at different dates, so it is expected that two imagery will have different cloud cover conditions, and finally, both of them can be merged together to create a cloudless image. 

This approach has been widely adopted in most satellite imagery processing software, but recently some GIS software also has this feature. ArcGIS desktop to be particular has a wide range of on-the-fly raster functions that can be used to perform this cloud removal task in a very fast and efficient way. Using raster functions implemented in Image Analysis Toolbar in ArcGIS, removing clouds in satellite imagery becomes easier than ever. If you are curious about how to do it, I already prepare a video tutorial that shows you how to do it below.