Graph and Total Variation Regularized Low-Rank Representation for Hyperspectral Anomaly Detection

Cheng, Tongkai; Wang, Bin

doi:10.1109/tgrs.2019.2936609

Cited by 134 publications

(55 citation statements)

References 50 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In most detection tasks, good performance always closely related to good semantic features, while combination of features always more discriminative than individual ones [17,18]. The pairwise graph constraint is introduced in [19] to ensure that the local intrinsic geometric structure of latent space be consistant with that in the original domain.…”

Section: Supergraphmentioning

confidence: 99%

“…where σ is a scalar parameter, and N k (x i ) is the set of knearest neighbors of x i . L = D − W is the graph Laplacian matrix, wherein degree matrix D is a diagonal matrix, and d ii is the summation of the i-th row of the symmetric matrix W. The graph regularization encourages samples within the same semantic region to share similar representation, and vise versa [17,20]. Whereas, there are two major disadvantages need addressing: 1) neither AE nor graph regularization takes the spatial information into consideration, which is vital in hyperspectral anomaly detection; 2) formulation of the graph Laplacian matrix L is time-consuming, it has to traverse the entire image to select the k-nearest neighbors for each PUT.…”

Section: Supergraphmentioning

confidence: 99%

See 1 more Smart Citation

Robust Graph Autoencoder for Hyperspectral Anomaly Detection

Fan

Huang

et al. 2021

ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

Autoencoder can not only extract features in an unsupervised manner, but also selects samples out that differs significantly from others. However, autoencoder is sensitive to noise and anomalies during training, and the relationships between pixels are discarded. In order to tackle these problems, we propose a robust graph autoencoder (RGAE) for hyperspectral anomaly detection. To be specific, we first redesign the objective function to encourage the network more robust to noise and anomalies. Meanwhile, a superpixel segmentation-based graph regularization term (SuperGraph) is incorporated into AE to preserve the geometric structure and spatial information simultaneously. Experiments with three real data sets are conducted to evaluate the performance, and the detection results demonstrate that our method outperforms other state-ofthe-art hyperspectral anomaly detectors.

show abstract

Section: Supergraphmentioning

confidence: 99%

Section: Supergraphmentioning

confidence: 99%

Robust Graph Autoencoder for Hyperspectral Anomaly Detection

Fan

Huang

et al. 2021

ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

show abstract

“…More recently, some efforts have been devoted to addressing the above-mentioned problems to some extent. In the anomaly detection task, several commonly used properties about the background pixels (low rank or spatially smooth) and anomalous targets (sparse) [19], [20] have also been applied in the target detection problem [21], in which a detector based on the sparse and low-rank matrix decomposition (SLRMD) of the observed data is established. However, the recovery of background is still unsatisfactory without the use of an appropriate background dictionary [22].…”

Section: Introductionmentioning

confidence: 99%

Decomposition Model With Background Dictionary Learning for Hyperspectral Target Detection

Cheng

Wang

2021

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

Self Cite

View full text Add to dashboard Cite

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, the background dictionary generated through the dual windows is susceptible to target contamination. Aiming to address these issues, in this paper, we propose a decomposition model (DM) with background dictionary learning (BDL) for hyperspectral target detection. The observed data are decomposed into three parts: background, target and noise. The background and target dictionaries are utilized to represent the background and target components, respectively. In order to achieve a satisfactory recovery of background and target components, the proposed DM exploits the spatial smoothness of background pixels and the scarcity of targets of interest in the whole scene via the total variation (TV) and the sparsity, respectively. Then, the separated target image is directly used for the detection purpose. Furthermore, a novel BDL method based on locality-constrained linear coding (LLC) is presented, and a complete and compact background dictionary can be learned with a low computational cost. Meanwhile, the a priori target dictionary is also incorporated into the learning process in order to suppress the contamination of target signal on the learned background spectra. Extensive experiments on both simulated and real hyperspectral data sets demonstrate the superiority of the proposed detector in comparison with several conventional and state-of-the-art target detectors.

show abstract

“…Thus, it is difficult to obtain the precise spectral information for the supervised method. However, unsupervised target detection methods, which are also referred to as anomaly detection methods and require no prior information, have more important research value in reality and have, thus, been widely applied in various fields, including food quality, safety control, search and rescue, mineral detection, and environmental surveillance [2,[20][21][22].…”

Section: Introductionmentioning

confidence: 99%

“…In addition, a number of representation-based methods have been proposed. These methods do not require any statistical assumptions and have, thus, attracted significant attention [20]. These techniques make use of the conspicuous characteristics of anomalies: the low probability of occurrence and the different spectral signature from the background pixels [27].…”

Section: Introductionmentioning

confidence: 99%

Random Collective Representation-Based Detector with Multiple Features for Hyperspectral Images

et al. 2021

Remote Sensing

View full text Add to dashboard Cite

Collaborative representation-based detector (CRD), as the most representative anomaly detection method, has been widely applied in the field of hyperspectral anomaly detection (HAD). However, the sliding dual window of the original CRD introduces high computational complexity. Moreover, most HAD models only consider a single spectral or spatial feature of the hyperspectral image (HSI), which is unhelpful for improving detection accuracy. To solve these problems, in terms of speed and accuracy, we propose a novel anomaly detection approach, named Random Collective Representation-based Detector with Multiple Feature (RCRDMF). This method includes the following steps. This method first extract the different features include spectral feature, Gabor feature, extended multiattribute profile (EMAP) feature, and extended morphological profile (EMP) feature matrix from the HSI image, which enables us to improve the accuracy of HAD by combining the multiple spectral and spatial features. The ensemble and random collaborative representation detector (ERCRD) method is then applied, which can improve the anomaly detection speed. Finally, an adaptive weight approach is proposed to calculate the weight for each feature. Experimental results on six hyperspectral datasets demonstrate that the proposed approach has the superiority over accuracy and speed.

show abstract

Graph and Total Variation Regularized Low-Rank Representation for Hyperspectral Anomaly Detection

Cited by 134 publications

References 50 publications

Robust Graph Autoencoder for Hyperspectral Anomaly Detection

Robust Graph Autoencoder for Hyperspectral Anomaly Detection

Decomposition Model With Background Dictionary Learning for Hyperspectral Target Detection

Random Collective Representation-Based Detector with Multiple Features for Hyperspectral Images

Contact Info

Product

Resources

About