Translator Disclaimer
3 November 2010 Multitemporal RADARSAT-2 polarimetric SAR data for urban land-cover mapping
Author Affiliations +
Proceedings Volume 7841, Sixth International Symposium on Digital Earth: Data Processing and Applications; 78410N (2010) https://doi.org/10.1117/12.873218
Event: The Sixth International Symposium on Digital Earth, 2009, Beijing, China
Abstract
The objective of this research is to evaluate the performance of multitemporal RADARSAT-2 polarimetric SAR data for urban land use/land-cover classification. Three dates of RADARSAT-2 polarimetric SAR data were acquired during the summer of 2008 over the rural-urban fringe of the Greater Toronto Area. The major land-cover types are residential areas, industry areas, bare land, golf courses, forest, and agricultural crops. The methodology used in this study follow the manner that first extracting the features and then carrying out the supervised classification taking the different feature combinations as an input. Support vectors machine is selected to be the classifier. SAR features including amplitude, intensity, long-term coherence, Freeman-Durden decomposition are extracted and compared by evaluating the classification abilities. Long-term coherence plays an important role in building discrimination in this study. The best classification results achieved by using the three dates HH, VH, HV amplitude layers and the coherence map. The overall accuracy is 82.3%. The results indicate that RADARSAT-2 polarimetric data has a potential to urban land-cover classification with the proper feature combinations.
© (2010) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Liang Gao and Yifang Ban "Multitemporal RADARSAT-2 polarimetric SAR data for urban land-cover mapping", Proc. SPIE 7841, Sixth International Symposium on Digital Earth: Data Processing and Applications, 78410N (3 November 2010); https://doi.org/10.1117/12.873218
PROCEEDINGS
9 PAGES


SHARE
Advertisement
Advertisement
Back to Top