Polarimetric SAR data ground object classification method based on multiple feature sets

The present invention discloses a polarimetric SAR data ground object classification method based on multiple feature sets. The polarimetric SAR data ground object classification method comprises the implementing steps of: (1) inputting data; (2) carrying out delicate Lee filtering; (3) extracting two feature sets; (4) carrying out clustering on the two feature sets; (5) comparing clustering results; (6) carrying out iteration classification; and (7) outputting a result. Compared with a Wishart classification method based on a single feature set in the prior art, the polarimetric SAR data ground object classification method based on the multiple feature sets has the advantages of improving classification accuracy of polarimetric synthetic aperture radar (SAR) data, reducing iterations when the polarimetric SAR data is classified by using the Wishart classification method and solving the problems of insufficiency for utilization of polarimetric information and a great number of wrongly classified samples in a training sample set for the Wishart classification method. The polarimetric SAR data ground object classification method based on the multiple feature sets can be applied to ground object classification of the polarimetric SAR data.