Some radio frequency and optical sensors collect large-scale sets of spatial imagery data whose content is often obscured by fog, clouds, foliage and other intervening structures. Often, the obstruction is such as to render unreliable the definition of underling images. There are several typical mathematical methods used in image processing to remove interferences from images to include spectral methods, wave front or shock methods, and the use of non-abelian group operations. In this paper, a new advanced image processing technique based on image segmentations has been developed and tested for the removal of fog, clouds, foliage and other interfering structures. The developed method has been applied to certain images to demonstrate its effectiveness in removing unwanted sub-images.