An iterative algorithm for removing the effect of thin cloud cover from LANDSAT imagery

In this paper, we propose an iterative algorithm for removing the effect of thin cloud cover from LANDSAT imagery. It is seen that the noise in such images is multiplicative as well as additive. The recorded image is first processed to update different parameters of the image formation model to known values. Processing algorithm and knowledge of parameter values are developed by considering the physics of the situation. A low-pass filter is then applied to the processed image to remove the effect of the cloud. The filter is of a tapered shape, and its parameters are adjusted to minimize the estimation error.