Cloud filtering for Landsat TM satellite images using multiple temporal mosaicing

Abstract

Cloud removal is a very important preprocessing step in remote sensing image analysis. In some remote sensing applications, a clean image free of cloud composite from a series of images taken in a short period of time will suffice for further analysis. This task is primarily carried out manually and time consuming. It is highly desirable to have some fully automated method to solve this problem efficiently. To this end, we propose a method called multiple temporal mosaicing as it is mimicking the process of mosacing pieces from images to obtain a whole image. We tested this method on Landsat TM 5 and 7 scenes. The resulting images show the effectiveness of this method clearly.