Abstract:

Passive microwave unmixing methods are very important for successful applications of passive microwave radiometer data. We proposed a passive microwave unmixing method based on land surface classification information. Through designing the simulation study areas with the component brightness temperature distributions based on different land surface types, we compared the proposed method with the method introduced by Bellerby. Furthermore, the two unmixing methods were evaluated and the main factors affecting the unmixing result were analyzed. Bellerby’s method is more suitable for the observation area where the land component brightness temperature was stable. When obvious change of the land component brightness temperature occurs in the observation area, the proposed method can obtain more high precision unmixing results. The research results demonstrated that the accuracy of the component brightness temperature using the proposed method was better than that using the Bellerby’s method. The precision of land component brightness temperature and water brightness temperature using the proposed method is approximately 0.1 K and 0.2 K respectively, while that using the Bellerby’s method is approximately 3.1 K and 5.9 K respectively.

Key words: image processing, component brightness temperature, passive microwave remote sensing data, error analysis, unmixing method