Segmentation of diverse image types using opening and closing

A flexible approach to image segmentation, originally developed for optical astronomical images, is useful for analyzing infrared oceanographic images and solar magnetograms; it should be applicable in other problem domains. The technique is based on the morphological operations of opening and closing. The original computer implementation was designed to find star boundaries in optical astronomical images, where the background is dark and stars, the brightest objects, are uniform or increase in intensity toward their centers. For infrared ocean images, the aim is to find fronts and associated rings; for magnetograms, the aim is to find sunspots, magnetically active regions of the solar surface. Rings, fronts, and sunspots are not very star-like. Hence these images must be pre-processed before applying the method. The edges found from images processed with the algorithm are useful as inputs to other processing, e.g. in automated tracking of ocean features