Segmentation of Optical Remote Sensing Images for Detecting Homogeneous Regions in Space and Time

Abstract

With the amount of multitemporal and multiresolution images growing exponentially, the number of image segmentation applications is recently increasing and, simultaneously, new challenges arise. Hence, there is a need to explore new segmentation concepts and techniques that make use of the data temporality. This study describes a spatio-temporal segmentation that adapts the traditional region growing technique to detect homogeneous regions in space and time in optical remote sensing images. Tests were conducted by considering the Dynamic Time Warping measure as the homogeneity criterion. Study cases on high temporal resolution for sequences of MODIS and Landsat-8 OLI vegetation indices products and comparisons with other distance measurements provided satisfactory outcomes.

Keywords

• Spatio-temporal segmentation
• Image Processing
• Dynamic Time Warping