Sea-Ice Thickness Inversion Of Multi-feature Fusion Based On C-band PolSAR Images

Polarimetric Synthetic Aperture Radar (PolSAR) not only has ability of all weather and day and night measurement, but can provide a wealth of polarimetric information, enabling it to be an ideal sea ice detection data source. Sea-Ice thickness is estimation of ice and the key indicators for ice disaster assessment and evaluation. Sea ice thickness is not only closely related to temperature changes but also the influence of surface roughness and the Spindrift, using remote sensing to sea ice thickness inversion is harder than larger. Studies have shown that:C-band in the sea-ice thickness inversion effect is significantly higher than that of L-band and X-band. Based on the background above, this paper using C-band polarimetric SAR data for inversion of sea ice thickness, and through the HEM data validation. Finally the dissertation presents the sea ice thickness inversion model based on fusion of multi level ice. Main work and findings are summarized as follows:First, polarization feature analysis and selection. In this paper, Firstly through the HEM and polarization characteristics of sea ice data analysis polarimetric scattering coefficient of sea ice and there is a certain relationship between sea ice thickness. Then high priority strategy based on the correlation coefficient for feature selection.Second, single feature inversion. Firstly, the single feature inversion model for screening out, and the sea ice thickness inversion is carried out by the single feature inversion model. Then, use the HEM data to verify the results. The results show that the correlation between the polarization features and the thickness is correct.Third, Sea-Ice thickness inversion of multi-feature fusion. Through simple linear regression method to select more than one feature to fusion, and figure out multiple features fusion based on C-band PolSAR inversion of sea-ice thickness model. And via the HEM data validation and feature inversion results comparing to the single feature inversion of sea-ice thickness inversion. The results show that the multi-feature inversion models better than the single feature inversion.Last, Multi-feature fusion model validation. Thickness inversion of three SIR-C data using the proposed model. The results show that the method of sea ice thickness inversion can effectively improve the accuracy of sea ice inversion, and get the effective inversion results.