A Novel Technique Based on Morphological Filters for Change Detection in Optical Remote Sensing Images

A novel technique for change detection based on the use of morphological filters is presented for very high geometrical resolution images. The comparison between two multitemporal images of the same zone is carried out with a standard change vector analysis technique. The obtained difference image is enhanced exploiting the non-linearity of the morphological operators. Different operators are analyzed and compared with respect to the change-detection problem. Alternating sequential filters by reconstruction proved to be the most effective, permitting to the preservation of the geometrical information of the structures in the scene while filtering the homogeneous areas. Experimental results confirm that the proposed technique increases the accuracy of the change detection with respect to the standard change vector analysis approach.