Edge detection technique on objects present in high resolution optical satellite images

Abstract

ABSTRACT Satellite images provide information on land cover, these will depend on the spatial resolution for analysis, in this paper we present a technique for image analysis based on edge detection using sobel-based detectors, in their different shapes. High-resolution satellite images provide a level of detail of land cover in the submetric range, which is equivalent to detecting objects that measure less than 1 meter on average, which is interpreted in one pixel in the image; With the analysis of high resolution images, objects such as ships can be determined, which are found in ports as well as on the high seas. Edge detectors applied to objects present in high resolution images help in the analysis of the morphology of the object. In the present study, high resolution images corresponding to boats were analyzed, which allows determining the contour of the boat with respect to sea water. The application of sobel-based edge detectors and their different shapes allowed determining a greater number of areas that corresponds to the perimeter of the ship. The results of the present work can be used in pattern recognition techniques to determine the characteristics of the objects to be analyzed.