Anomaly detection is of great importance among hyperspectral applications, which aims at locating targets that are spectrally different from their surrounding background. A variety anomaly methods have been proposed in the past. However, most them fail to take high spectral correlations all pixels into consideration. Low-rank representation (LRR) has drawn a deal interest recent years, as promising model exploit intrinsic low-rank property images. Nevertheless, original LRR only analyzes...

Anomaly detection in hyperspectral imagery has been an active topic among the remote sensing applications. It aims at identifying anomalous targets with different spectra from their surrounding background. Therefore, effective detector should be able to distinguish anomalies, especially for weak ones, background and noise. In this article, we propose a novel method anomaly based on total variation (TV) sparsity regularized decomposition model. This model decomposes into three components:...

Representation-based target detectors for hyperspectral imagery have attracted considerable attention in recent years. However, their detection performance is still unsatisfactory due to the independent manner of recovery process on each test pixel. Moreover, background dictionary generated through dual windows susceptible contamination. Aiming address these issues, this article, we propose a decomposition model (DM) with learning (BDL) detection. The observed data are decomposed into three...

Anomaly detection is an important task among hyperspectral applications, which detects the anomalous targets that are spectrally different from their surroundings. Most of anomaly detectors construct models in linear space, where nonlinear characteristics data not taken into consideration. The kernel-based methods, implicitly map a high-dimensional feature have shown great potential dealing with problems. Nevertheless, original methods only exploit spectral features without taking advantage...

Anomaly detection has been known to be an important issue in hyperspectral remote sensing applications. It aims detect anomalous targets whose spectral signatures are very different from the background pixels. Although many linear detectors have obtained acceptable results, model might not able describe complex data and could replaced by nonlinear models. In this article, we investigate intrinsic characteristics of images (HSIs) on basis mixing models propose a novel anomaly method based...

Improving the performance of nonlinear unmixing has become an active topic among remote sensing applications. Usually, noise levels hyperspectral images (HSIs) vary with different bands. However, this fact is generally ignored and may, to some extent, result in a degradation results. Nonetheless, valuable spatial information that provides great potential for improving seldom been considered current unmixing. In letter, we propose novel kernel-based model which band-wise characterization...

A novel method for hyperspectral anomaly detection based on low-rank representation with manifold regularization is proposed in this paper. Usually, a imagery can be modeled as superposition of two parts: background part low rank dimensionality and described by sparse matrix. Low-rank (LRR) used to find the lowest all pixels jointly which represents part, then contained residual original image. To learn more discriminative representation, we incorporate term into LRR model. An important...

This paper presents a novel method for hyperspectral anomaly detection based on total variation and structured dictionary. Generally, imagery can be modeled as superposition of two components: background anomalies. Since each pixel in the well represented by some other pixels anomalies to detected approximately potential anomalous pixels. Therefore, test using dictionary consisting Moreover, considering spatial homogeneity natural sparse nature anomalies, regularization terms named sparisty...