Remote sensing image classification method based on high-dimensional feature selection and multi-level fusion

The invention discloses a remote sensing image classification method based on high-dimensional feature selection and multi-level fusion. The method includes the following steps: firstly, on the basisof analyzing a special imaging mechanism of remote sensing images, extracting various heterogeneous features such as the shape, spectrum and texture of the remote sensing images; secondly, for the extracted high-dimensional heterogeneous features, on the one hand, performing feature selection on various high-dimensional features by using a class information sub-set feature selection algorithm to obtain a corresponding group of optimal low-dimensional feature sub-sets; and on the other hand, fusing every two types of high-dimensional heterogeneous features by using a discriminant correlation analysis algorithm to obtain corresponding low-dimensional fusion features; thirdly, respectively inputting each optimal low-dimensional feature sub-set and the low-dimensional fusion features into an SVM classifier, and initia