Clutter suppression based on adaptive subspace reconstruction
基于自适应子空间重建的杂波抑制

To suppress clutter in infrared image, a novel clutter suppression method based on adaptive subspace construction is proposed. An over-complete dictionary for modeling target is learned based on the sparse coding theory at first. Then the image patch is extracted from the test image in order . Representation coefficients are computed according to the dictionary. According to the difference in the representation coefficients of the target patch and the background patch in sparse domain, adaptive subspace is generated to reconstruct the image patch. Thus a residual image between original image patch and reconstructed patch is obtained. The experiment results show that the residual image obtained by the proposed method can effectively suppress background clutter and significantly improve signal-to-noise ratio of the infrared image.