Pregled algoritmov in programske opreme za segmentacijo optičnih daljinsko zaznanih posnetkov ( = Overview of segmentation algorithms and software for optical remote sensing imagery)

Abstract

Sentinel satellites and the European Copernicus programme have begun to provide huge amounts of high spatial and temporal resolution images, which need to be processed quickly and preferably automatically. Object based classification is the most commonly used method for the interpretation of high spatial resolution optical images. It is a challenging process of object borders recognition (segmentation) and classification of segments in the user-defined classes (classification). The quality of the segmentation is crucial because it has a direct impact on the quality of final classification results. Several algorithms have been developed for optical remote sensing imagery segmentation with varying (depending on the purpose of the analysis) accuracy. Knowledge of algorithms and adequacy of application are crucial for high quality segmentation results. In this paper, we give an overview of optical remote sensing imagery segmentation algorithms and the possibility of their application. Besides, an overview of object based classification software, with a focus on segmentation process, is given. The intent of this paper is not to provide a recipe for the best algorithm and software selection, but to support users with knowledge they might need to be able to obtain the best possible segmentation results of an analysis in given circumstances.

Keywords

• remote sensing
• segmentation
• object based classification