Hyperspectral anomaly detection:a performance comparison of existing techniques

Anomaly detection in Hyperspectral Imagery(HSI)has received considerable attention because of its potential application in several areas.Numerous anomaly detection algorithms for HSI have been proposed in the literature;however,due to the use of different datasets in previous studies,an extensive performance comparison of these algorithms is missing.In this paper,an overview of the current state of research in hyperspectral anomaly detection is presented by broadly dividing all the previously proposed algorithms into eight different categories.In addition,this paper presents the most comprehensive comparative analysis to-date in hyperspectral anomaly detection by evaluating 22 algorithms on 17 different publicly available datasets.Results indicate that attribute and edge-preserving filtering-based detection(AED),local summation anomaly detection based on collaborative representation and inverse distance weight(LSAD-CR-IDW)and local summation unsupervised nearest regularized subspace with an outlier removal anomaly detector(LSUNRSORAD)perform better as indicated by the mean and median values of area under the receiver operating characteristic(ROC)curves.Finally,this paper studies the effect of various dimensionality reduction techniques on anomaly detection.Results indicate that reducing the number of components to around 20 improves the performance;however,any further decrease deteriorates the performance.

Anomaly detection of hot components in gas turbine based on frequent pattern extraction

Hot components operate in a high-temperature and high-pressure environment. The occurrence of a fault in hot components leads to high economic losses. In general, exhaust gas temperature(EGT) is used to monitor the performance of hot components.However, during the early stages of a failure, the fault information is weak, and is simultaneously affected by various types of interference, such as the complex working conditions, ambient conditions, gradual performance degradation of the compressors and turbines, and noise. Additionally, inadequate effective information of the gas turbine also restricts the establishment of the detection model. To solve the above problems, this paper proposes an anomaly detection method based on frequent pattern extraction. A frequent pattern model(FPM) is applied to indicate the inherent regularity of change in EGT occurring from different types of interference. In this study, based on a genetic algorithm and support vector machine regression, the relationship model between the EGT and interference was tentatively built. The modeling accuracy was then further improved through the selection of the kernel function and training data. Experiments indicate that the optimal kernel function is linear and that the optimal training data should be balanced in addition to covering the appropriate range of operating conditions and ambient temperature. Furthermore, the thresholds based on the Pauta criterion that is automatically obtained during the modeling process, are used to determine whether hot components are operating abnormally. Moreover, the FPM is compared with the similarity theory, which demonstrates that the FPM can better suppress the effect of the component performance degradation and fuel heat value fluctuation. Finally, the effectiveness of the proposed method is validated on seven months of actual data obtained from a Titan130 gas turbine on an offshore oil platform. The results indicate that the proposed method can sensitively detect malfunctions in hot components during the early stages of a fault, and is robust to various types of interference.