SAR ATR based on multi-subspaces of independent component analysis

In this paper, a SAR (synthetic aperture radar) ATR algorithm based on multi-subspaces of Independent Component Analysis (ICA) is proposed. Features of each target SAR images are extracted by using ICA firstly, and then the Feature images of the target compose a subspace. All targets can compose multi-subspaces. Finally, we combine all the subspaces to a big new subspace by using PCA. We test our method on MASTAR database. The experiment result shows that the performance of this method is superior to other methods.