Assessing the quality of DSM from ALOS optical and radar data for automatic drainage extraction

Abstract

The hydrological network analysis was based for many years on topographic maps or on aerial photos. The evolution of GIS software and the existence of plenty of satellite sensors collecting data with high spatial resolution made possible the creation of quite accurate Digital Surface Models (DSM) and the automatic drainage extraction. In this study the suitability and the accuracy of drainage network derived from ALOS optical and radar data are validated with reference to the respective information from the topographic maps of 1/50.000. DSM from ALOS PALSAR and ALOS PRISM are created, evaluated for both vertical and horizontal accuracy and used for automatic drainage extraction. A fifth order basin of Alfios River in Western Peloponese, Greece was selected for the validation. Hydrological parameters such as the stream length, the drainage frequency and the drainage density were used for the evaluation of seven DSMs. It was proved that the DSMs of the ALOS optical data are more suitable for hydrological analysis and that the vertical high accuracy of the DSM doesn’t guarantee its suitability for automatic drainage extraction.