A combined weighted multilevel morphological filter for edge detection postprocessing

In this paper, based on the basic morphological filters and correlation theory, a combined weighted multilevel morphological filter is proposed for smoothing the rough edges of the target detected from an electro-optical image. In computer simulation, the results show that the filter not only improves the boundaries of the target but also saves its characteristics. This work lays the foundation of target detection and recognition of electro-optical trackers.