Compact polarimetry assessment for rice and wetland mapping

Abstract

Polarimetric RADARSAT-2 data of rice and wetlands are used to simulate compact polarimetry (CP) mode data from the upcoming RADARSAT Constellation Mission (RCM). The simulated CP data are then used to evaluate the information content for rice and wetland mapping using supervised classification, and the results are compared for linear and circular polarization combinations and polarimetric decompositions from the fully polarimetric data and the simulated CP data. The results are consistent for both rice and wetlands and show that the classification accuracy increases as one goes up the polarization hierarchy. The circular polarizations produced the best classification results for the polarization combinations. This result requires further research to verify. Although the CP data did not perform as well as the fully polarimetric data, the results were better than for dual polarization, and this mode may offer the best option for rice and wetland mapping applications because of swath coverage. Note that both the compact simulations and the fully polarimetric data produced operationally suitable classification accuracies. Additional research is underway to evaluate the monitoring capability of this new CP mode. This article describes the approach used for the analyses and the classification results for both rice and wetlands.

Acknowledgements

This project was supported by the Remote Sensing Science programme at CCRS at the Earth Sciences Sector Natural resources Canada and by the Canadian Space Agency through the RCM programme at CCRS and the SOAR programme at IRSA. Data processing support from Vince Decker at CCRS is greatly appreciated. Finally, thanks to Keith Raney for ongoing support to the compact polarimetry project at CCRS. ESS Contribution number/Numéro de contribution du SST: 20110301.