Urban Monitoring at a Regional Scale Based on MERIS and ASAR data
Luis Gomez-Chova(1) and Diego Fernández-Prieto(2)
University of Valencia, GPDS,
Dr. Moliner 50, Ing. Electronica,
E-46100, Burjassot (Valencia),
(2) European Space Agency, ESA-ESRIN, Via Galileo Galilei, EO Science & Applications, 00044, Frascati (Rome), Italy
Monitoring urban areas at a regional scale, and even at a global scale, has become an increasingly important topic in the last decades. New Earth Observation (EO) platforms provide remotely-sensed imagery that allows the monitoring of urban expansion with the required spatial and temporal resolution. Currently, we can take advantage of the combination of sensors present in the ESA ENVISAT platform which offers simultaneous acquisition of MERIS (MEdium Resolution Imaging Spectrometer) and ASAR (Advanced Synthetic Aperture Radar) sensors.
The information retrieved by optical and SAR sensors differs to a great extent one from another (spectral, structural and temporal features), and pattern recognition techniques have demonstrated excellent capabilities for detecting the urban tissue accurately when combining multispectral and SAR data. The aim of this work is to demonstrate the capabilities of MERIS and ASAR data to map urban areas at a regional scale. Since MERIS FR works with a 300m pixel size, this application focuses on obtaining high accuracy at a regional scale, providing the basis for a low-cost highly automated pan-European service dedicated to identifying and monitoring urban areas, for instance, of EU25. Therefore, it is not interesting to obtain an exhaustive map with all the thematic classes present in an area, but to produce an automatic classification of ’Urban/Non-Urban’. This binary information is of great use for European policy bodies such EEA or DG-Regio in different aspects such as environmental monitoring, soil protection and regional cohesion policies and spatial planning.
For this purpose, a two-stage classification scheme based on an unsupervised approach is proposed, which allows us to introduce supervised information about the class of interest without an additional sample labelling. Results of this work were obtained under framework of the ESA Category-1 project (C1P-ID2489) titled “Development of an Specialized Classification System for Urban Monitoring at Regional Scale Based on ASAR and MERIS data”. Full details will be shown at the time of the conference.