Polarimetric SAR (PolSAR) image classification is one of the important applications of PolSAR remote sensing. It is a difficult high-dimension nonlinear mapping problem, the sparse representations based on learning overcomplete dictionary have shown great potential to solve such problem. The overcomplete dictionary plays an important role in PolSAR image classification, however for PolSAR image complex scenes, features shared by different classes will weaken the discrimination of learned dictionary, so as to degrade classification performance. In this paper, we propose a novel overcomplete dictionary learning model to enhance the discrimination of dictionary. The learned overcomplete dictionary by the proposed model is more discriminative and very suitable for PolSAR classification.
Proc. SPIE. 9811, MIPPR 2015: Multispectral Image Acquisition, Processing, and Analysis
KEYWORDS: Signal to noise ratio, Principal component analysis, Image compression, Chemical species, Denoising, Interference (communication), Image quality, Associative arrays, Image denoising, Signal analyzers
This article addresses the image denoising problem in the situations of strong noise. The method we propose is intended to preserve faint signal details under these difficult circumstances. The new method we introduce, called principal basis analysis, is based on a novel criterion: the reproducibility which is an intrinsic characteristic of the geometric regularity in natural images. We show how to measure reproducibility. Then we present the principal basis analysis method, which chooses, in sparse representation of the signal, the components optimizing the reproducibility degree to build a so-called principal basis. With this principal basis, we show that a noise-free reconstruction may be obtained. As illustrations, we apply the principal signal basis to image denoising for natural images with details in low signal-to-noise ratio, showing performance better than some reference methods.