Analysis of Microwave Radiometry of Snow and Ice on an Outdoor Experimental Asphalt Surface

This letter presents the feasibility of a classification that divides six surfaces into more than four surfaces using radiometric values retrieved from brightness temperatures (TBs) observed at 6- and 36-GHz radiometers with vertical (V) and horizontal (H) polarizations on asphalt surface. The feasibility was investigated by using the Mahalanobis-distance-based approach of the canonical discriminant analysis. Combining 6 V (or 36 V) and 36 H emissivities, with the use of the surface temperature, showed the classification accuracy of 94%. In addition, combining the polarization ratio at 36 GHz TBs (PR<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">36</sub>) and the cross-polarized gradient ratio between 36 H and 6 V TBs (XGPR<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">36H06V</sub>), without the use of the surface temperature, showed the classification accuracy of 97%. Both the combination of 6 V and 36 H emissivities and the combination of PR<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">36</sub> and XGPR<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">36H06V</sub> have the potential for dividing six surface conditions into five surface conditions. Results suggest that the combination of PR<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">36</sub> and XGPR<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">36H06V</sub> potentially has a classification ability similar to that of 6 V and 36 H emissivities.