Hyperspectral Anomaly Detection Using Collaborative Representation With Outlier Removal

Su, Hongjun; Wu, Zhaoyue; Du, Qian; Du, Peijun

doi:10.1109/jstars.2018.2880749

Cited by 80 publications

(31 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the recent years, new approaches to the problem have emerged in order to cope, for instance, with inadequate Gaussian-distributed representations for non-homogeneous backgrounds [24], the presence of noise in the images, by using a combined similarity criterion anomaly detector (CSCAD) method [25], as well as the removal of outliers by using a collaborative representation detector (CRD) [26,27].…”

Section: Anomaly Detection Algorithmsmentioning

confidence: 99%

A Simulation Environment for Validation and Verification of Real Time Hyperspectral Processing Algorithms on-Board a UAV

et al. 2019

View full text Add to dashboard Cite

The utilization of hyperspectral imaging sensors has gained a significant relevance among many different applications due to their capability for collecting a huge amount of information across the electromagnetic spectrum. These sensors have been traditionally mounted on-board satellites and airplanes in order to extract information from the Earth’s surface. Fortunately, the progressive miniaturization of these sensors during the last lustrum has enabled their use in other remote sensing platforms, such as drones equipped with hyperspectral cameras which bring advantages in terms of higher spatial resolution of the acquired images, more flexible revisit times and lower cost of the flight campaigns. However, when these drones are autonomously flying and taking real-time critical decisions from the information contained in the captured images, it is crucial that the whole process takes place in a safe and predictable manner. In order to deal with this problem, a simulation environment is presented in this work to analyze the virtual behavior of a drone equipped with a pushbroom hyperspectral camera used for assisting harvesting applications, which enables an exhaustive and realistic validation and verification of the drone real-time hyperspectral imaging system prior to its launch. To the best of the authors’ knowledge, the proposed environment represents the only solution in the state-of-the-art that allows the virtual verification of real-time hyperspectral image processing algorithms under realistic conditions.

show abstract

Section: Anomaly Detection Algorithmsmentioning

confidence: 99%

A Simulation Environment for Validation and Verification of Real Time Hyperspectral Processing Algorithms on-Board a UAV

et al. 2019

View full text Add to dashboard Cite

show abstract

“…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]. There are several forms of representation-based methods, including the sparse representation-based methods [1,[31][32][33][34][35][36][37], the low-rank methods [38][39][40][41][42][43], and the collaborative methods [44][45][46][47][48].…”

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

“…orthogonal to the background subspace, and can be detected by maximizing the signal-to-noise (SNR) ratio in the subspace orthogonal to the background subspace. In recent years, sparse representation (SR) has been emerging as an efficient methodology, and has been successfully applied for many computer vision applications [38][39][40][41][42][43][44][45][46][47][48][49][50]. With SR theory, the Sparse Representation-based target Detector (SRD) has been developed, which first represents a test pixel using the union of the background and target dictionaries via sparsity-inducing algorithms, such as the orthogonal matching pursuit (OMP) [17], and the test pixel is classified by comparing the residuals between the input test pixel with the two test pixels respectively reconstructed by the target and background dictionaries.…”

Section: Introductionmentioning

confidence: 99%

Learning Structurally Incoherent Background and Target Dictionaries for Hyperspectral Target Detection

Guo

Luo

Zhang

et al. 2020

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

View full text Add to dashboard Cite

Existing sparsity-based hyperspectral image (HSI) target detection methods have two key problems, i.e., 1) The background dictionary is locally constructed by the pixels between the inner and outer windows, surrounding and enclosing the central test pixel. The dual-window strategy is intricate, and might result in impure background dictionary deteriorating the detection performance. 2) For an unbalanced binary classification problem, the target dictionary atoms are generally inadequate comparing with the background dictionary, which might yield unstable performance. For the issues, this paper proposes a novel Structurally Incoherent Background and Target Dictionaries (SIBTD) learning model for HSI target detection. Specifically, with the concept that the observed HSI data is composed of lowrank background, sparsely distributed targets and dense Gaussian noise, the background and target dictionaries can be jointly derived from the observed HSI data. Additionally, the introduction of structural incoherence can enhance the discrimination between the target and background dictionaries. Thus, the developed model can not only lead to a pure and unified background dictionary, but also augment the target dictionary for improved detection performance. Besides, an efficient optimization algorithm is devised to solve SIBTD model, and the performance of SIBTD is verified on three benchmark HSI datasets in comparison with several state-of-the-art detectors.

show abstract

Hyperspectral Anomaly Detection Using Collaborative Representation With Outlier Removal

Cited by 80 publications

References 40 publications

A Simulation Environment for Validation and Verification of Real Time Hyperspectral Processing Algorithms on-Board a UAV

A Simulation Environment for Validation and Verification of Real Time Hyperspectral Processing Algorithms on-Board a UAV

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

Learning Structurally Incoherent Background and Target Dictionaries for Hyperspectral Target Detection

Contact Info

Product

Resources

About