Erosion After Gradient (ErAGrad) morphological profile

Morphological profiles [1, 2] have been proven useful for classification and segmentation of remote sensed images. A morphological profile contains information about the size and shape of objects in an image. However, the profile of a pixel not only depends on the geometry of the object it belongs to, but also on whether it is a bright, dark or gray object. Moreover, morphological profiles are not easily extendable to color images. Different attempts have been made, but they either require an artificial ordening of the colors or create profiles for every color band. This makes it more difficult to interpret the morphological profile. In this paper we propose a new kind of morphological profile, the ErAGrad profile. This profile contains a single value for each pixel on each scale. The ErA-Grad profile is easily interpretable and does not depend on the color of the object or the color of the background, but only on the difference between both.