The rapid integration of Internet of Things(IoT)technologies is reshaping the global energy landscape by deploying smart meters that enable high-resolution consumption monitoring,two-way communication,and advanced met...The rapid integration of Internet of Things(IoT)technologies is reshaping the global energy landscape by deploying smart meters that enable high-resolution consumption monitoring,two-way communication,and advanced metering infrastructure services.However,this digital transformation also exposes power system to evolving threats,ranging from cyber intrusions and electricity theft to device malfunctions,and the unpredictable nature of these anomalies,coupled with the scarcity of labeled fault data,makes realtime detection exceptionally challenging.To address these difficulties,a real-time decision support framework is presented for smart meter anomality detection that leverages rolling time windows and two self-supervised contrastive learning modules.The first module synthesizes diverse negative samples to overcome the lack of labeled anomalies,while the second captures intrinsic temporal patterns for enhanced contextual discrimination.The end-to-end framework continuously updates its model with rolling updated meter data to deliver timely identification of emerging abnormal behaviors in evolving grids.Extensive evaluations on eight publicly available smart meter datasets over seven diverse abnormal patterns testing demonstrate the effectiveness of the proposed full framework,achieving average recall and F1 score of more than 0.85.展开更多
Research on the range anomaly suppression algorithm in laser radar （ladar） range images is significant in the application and development of ladar. But most of existing algorithms cannot protect the edge and linear ...Research on the range anomaly suppression algorithm in laser radar （ladar） range images is significant in the application and development of ladar. But most of existing algorithms cannot protect the edge and linear target well while suppressing the range anomaly. Aiming at this problem, the differences among the edge, linear target, and range anomaly are analyzed and a novel algo- rithm based on neighborhood pixels detection is proposed. Firstly, the range differences between current pixel and its neighborhood pixels are calculated. Then, the number of neighborhood pixels is detected by the range difference threshold. Finally, whether the current pixel is a range anomaly is distinguished by the neighbor- hood pixel number threshold. Experimental results show that the new algorithm not only has a better range anomaly suppression performance and higher efficiency, but also protects the edge and linear target preferably compared with other algorithms.展开更多
For anomaly detection,anomalies existing in the background will affect the detection performance.Accordingly,a background refinement method based on the local density is proposed to remove the anomalies from thebackgr...For anomaly detection,anomalies existing in the background will affect the detection performance.Accordingly,a background refinement method based on the local density is proposed to remove the anomalies from thebackground.In this work,the local density is measured by its spectral neighbors through a certain radius which is obtained by calculating the mean median of the distance matrix.Further,a two-step segmentation strategy is designed.The first segmentation step divides the original background into two subsets,a large subset composed by background pixels and a small subset containing both background pixels and anomalies.The second segmentation step employing Otsu method with an aim to obtain a discrimination threshold is conducted on the small subset.Then the pixels whose local densities are lower than the threshold are removed.Finally,to validate the effectiveness of the proposed method,it combines Reed-Xiaoli detector and collaborative-representation-based detector to detect anomalies.Experiments are conducted on two real hyperspectral datasets.Results show that the proposed method achieves better detection performance.展开更多
Distinguishing geochemical anomalies from background is a basic task in exploratory geochemistry. The derivation of geochemical anomalies from stream sediment geochemical data and the decomposition of these anomalies ...Distinguishing geochemical anomalies from background is a basic task in exploratory geochemistry. The derivation of geochemical anomalies from stream sediment geochemical data and the decomposition of these anomalies into their component patterns were described. A set of stream sediment geochemical data was obtained for 1 880 km 2 of the Pangxidong area, which is in the southern part of the recently recognized Qinzhou-Hangzhou joint tectonic belt. This belt crosses southern China and tends to the northwest (NE) direction. The total number of collected samples was 7 236, and the concentrations of Ag, Au, Cu, As, Pb and Zn were measured for each sample. The spatial combination distribution law of geochemical elements and principal component analysis (PCA) were used to construct combination models for the identification of combinations of geochemical anomalies. Spectrum-area (S-A) fractal modeling was used to strengthen weak anomalies and separate them from the background. Composite anomaly modeling was combined with fractal filtering techniques to process and analyze the geochemical data. The raster maps of Au, Ag, Cu, As, Pb and Zn were obtained by the multifractal inverse distance weighted (MIDW) method. PCA was used to combine the Au, Ag, Cu, As, Pb, and Zn concentration values. The S-A fractal method was used to decompose the first component pattern achieved by the PCA. The results show that combination anomalies from a combination of variables coincide with the known mineralization of the study area. Although the combination anomalies cannot reflect local anomalies closely enough, high-anomaly areas indicate good sites for further exploration for unknown deposits. On this basis, anomaly and background separation from combination anomalies using fractal filtering techniques can provide guidance for later work.展开更多
It is difficult to detect the anomalies whose matching relationship among some data attributes is very different from others’ in a dataset. Aiming at this problem, an approach based on wavelet analysis for detecting ...It is difficult to detect the anomalies whose matching relationship among some data attributes is very different from others’ in a dataset. Aiming at this problem, an approach based on wavelet analysis for detecting and amending anomalous samples was proposed. Taking full advantage of wavelet analysis’ properties of multi-resolution and local analysis, this approach is able to detect and amend anomalous samples effectively. To realize the rapid numeric computation of wavelet translation for a discrete sequence, a modified algorithm based on Newton-Cores formula was also proposed. The experimental result shows that the approach is feasible with good result and good practicality.展开更多
Solar arrays are important and indispensable parts of spacecraft and provide energy support for spacecraft to operate in orbit and complete on-orbit missions.When a spacecraft is in orbit,because the solar array is ex...Solar arrays are important and indispensable parts of spacecraft and provide energy support for spacecraft to operate in orbit and complete on-orbit missions.When a spacecraft is in orbit,because the solar array is exposed to the harsh space environment,with increasing working time,the performance of its internal electronic components gradually degrade until abnormal damage occurs.This damage makes solar array power generation unable to fully meet the energy demand of a spacecraft.Therefore,timely and accurate detection of solar array anomalies is of great significance for the on-orbit operation and maintenance management of spacecraft.In this paper,we propose an anomaly detection method for spacecraft solar arrays based on the integrated least squares support vector machine(ILS-SVM)model:it selects correlated telemetry data from spacecraft solar arrays to form a training set and extracts n groups of training subsets from this set,then gets n corresponding least squares support vector machine(LS-SVM)submodels by training on these training subsets,respectively;after that,the ILS-SVM model is obtained by integrating these submodels through a weighting operation to increase the prediction accuracy and so on;finally,based on the obtained ILS-SVM model,a parameterfree and unsupervised anomaly determination method is proposed to detect the health status of solar arrays.We use the telemetry data set from a satellite in orbit to carry out experimental verification and find that the proposed method can diagnose solar array anomalies in time and can capture the signs before a solar array anomaly occurs,which reflects the applicability of the method.展开更多
Quantitative descriptions of geochemical patterns and providing geochemical anomaly map are important in applied geochemistry. Several statistical methodologies are presented in order to identify and separate geochemi...Quantitative descriptions of geochemical patterns and providing geochemical anomaly map are important in applied geochemistry. Several statistical methodologies are presented in order to identify and separate geochemical anomalies. The U-statistic method is one of the most important structural methods and is a kind of weighted mean that surrounding points of samples are considered in U value determination. However, it is able to separate the different anomalies based on only one variable. The main aim of the presented study is development of this method in a multivariate mode. For this purpose, U-statistic method should be combined with a multivariate method which devotes a new value to each sample based on several variables. Therefore, at the first step, the optimum p is calculated in p-norm distance and then U-statistic method is applied on p-norm distance values of the samples because p-norm distance is calculated based on several variables. This method is a combination of efficient U-statistic method and p-norm distance and is used for the first time in this research. Results show that p-norm distance of p=2(Euclidean distance) in the case of a fact that Au and As can be considered optimized p-norm distance with the lowest error. The samples indicated by the combination of these methods as anomalous are more regular, less dispersed and more accurate than using just the U-statistic or other nonstructural methods such as Mahalanobis distance. Also it was observed that the combination results are closely associated with the defined Au ore indication within the studied area. Finally, univariate and bivariate geochemical anomaly maps are provided for Au and As, which have been respectively prepared using U-statistic and its combination with Euclidean distance method.展开更多
This paper reports statistical results of Seismo-Ionospheric Anomalies(SIAs) of the Total Electron Content(TEC) in the Global Ionosphere Map(GIM) associated with 56 M≥6.0 earthquakes in China during 1998—2012.To det...This paper reports statistical results of Seismo-Ionospheric Anomalies(SIAs) of the Total Electron Content(TEC) in the Global Ionosphere Map(GIM) associated with 56 M≥6.0 earthquakes in China during 1998—2012.To detect SIA,a quartile-based(i.e.median-based) process is performed.TEC anomalies for the period of earthquakes without being led by magnetic storms about 10 days are further isolated and examined to confirm the SIP existence.Results show that SIA is the TEC significantly decrease in the afternoon period 2—9 days before the earthquakes in China,which is in a good agreement with the SIA appearing before the 12 May 2008 M 8.0 Wenchuan earthquake.展开更多
From Octobet 1998 to January 1999,5 earthquakes ( M s≥5) occurred between Ninglang and Yanyuan counties (27°07′～27°12′N,100°40′～101°00′E area).They were situated in 140km southwest of the Xi...From Octobet 1998 to January 1999,5 earthquakes ( M s≥5) occurred between Ninglang and Yanyuan counties (27°07′～27°12′N,100°40′～101°00′E area).They were situated in 140km southwest of the Xichang.Among them,the largest one is M s 6 2 on November 19,1998.Based on small seismic data by the seismic remote sensing station of Xichang and the seismological station of Muli,and regional observation data,passing through careful observation and scientific analyses,we had made better forecasts before the earthquakes.That results obvious social benefits.By processing data of precursory earthquakes,such as,original observation data of total geomagnetic intensity from the station of Xichang,pressure capacitance stressometer and quartz horizaontal pendulum tiltmeter from the Xiaomiao station of Xichang,we summarized the sequence characteristics of the series earthquakes.The information about short\|term anomaly of gruond strain,total geomagnetic intensity and ground tilt before the earthquake is emphatically explained.展开更多
In recent years,anomaly detection has attracted much attention in industrial production.As traditional anomaly detection methods usually rely on direct comparison of samples,they often ignore the intrinsic relationshi...In recent years,anomaly detection has attracted much attention in industrial production.As traditional anomaly detection methods usually rely on direct comparison of samples,they often ignore the intrinsic relationship between samples,resulting in poor accuracy in recognizing anomalous samples.To address this problem,a knowledge distillation anomaly detection method based on feature reconstruction was proposed in this study.Knowledge distillation was performed after inverting the structure of the teacher-student network to avoid the teacher-student network sharing the same inputs and similar structure.Representability was improved by using feature splicing to unify features at different levels,and the merged features were processed and reconstructed using an improved Transformer.The experimental results show that the proposed method achieves better performance on the MVTec dataset,verifying its effectiveness and feasibility in anomaly detection tasks.This study provides a new idea to improve the accuracy and efficiency of anomaly detection.展开更多
The anomaly detection of electromagnetic environment situation(EMES) has essential reference value for electromagnetic equipment behavior cognition and battlefield threat assessment.In this paper,we proposed a deep le...The anomaly detection of electromagnetic environment situation(EMES) has essential reference value for electromagnetic equipment behavior cognition and battlefield threat assessment.In this paper,we proposed a deep learning-based method for detecting anomalies in EMES to address the problem of relatively low efficiency of electromagnetic environment situation anomaly detection(EMES-AD).Firstly,the convolutional kernel extracts the static features of different regions of the EMES.Secondly,the dynamic features of the region are obtained by using a recurrent neural network(LSTM).Thirdly,the Spatio-temporal features of the region are recovered by using a de-convolutional network and then fused to predict the EMES.The structural similarity algorithm(SSIM) is used to determine whether it is anomalous.We developed the detection framework,de-signed the network parameters,simulated the data sets containing different anomalous types of EMES,and carried out the detection experiments.The experimental results show that the proposed method is effective.展开更多
图像异常检测旨在识别并定位图像中的异常区域,针对现有算法中不同层次特征信息利用不充分的问题,提出了基于多层次特征融合网络的图像异常检测算法。通过使用融合了异常先验知识的伪异常数据生成算法,对训练集进行了异常数据扩充,将异...图像异常检测旨在识别并定位图像中的异常区域,针对现有算法中不同层次特征信息利用不充分的问题,提出了基于多层次特征融合网络的图像异常检测算法。通过使用融合了异常先验知识的伪异常数据生成算法,对训练集进行了异常数据扩充,将异常检测任务转化为监督学习任务;构建了多层次特征融合网络,将神经网络中不同层次特征进行融合,丰富了特征中的低层纹理信息和高层语义信息,使得用于异常检测的特征更具区分性;训练时,设计了分数约束损失和一致性约束损失,并结合特征约束损失对整个网络模型进行训练。实验结果表明,MVTec数据集上图像级检测接收机工作特性曲线下面积(area under the receiver operating characteristic, AUROC)平均值为98.7%,像素级定位AUROC平均值为97.9%,每区域重叠率平均值为94.2%,均高于现有的异常检测算法。展开更多
文摘The rapid integration of Internet of Things(IoT)technologies is reshaping the global energy landscape by deploying smart meters that enable high-resolution consumption monitoring,two-way communication,and advanced metering infrastructure services.However,this digital transformation also exposes power system to evolving threats,ranging from cyber intrusions and electricity theft to device malfunctions,and the unpredictable nature of these anomalies,coupled with the scarcity of labeled fault data,makes realtime detection exceptionally challenging.To address these difficulties,a real-time decision support framework is presented for smart meter anomality detection that leverages rolling time windows and two self-supervised contrastive learning modules.The first module synthesizes diverse negative samples to overcome the lack of labeled anomalies,while the second captures intrinsic temporal patterns for enhanced contextual discrimination.The end-to-end framework continuously updates its model with rolling updated meter data to deliver timely identification of emerging abnormal behaviors in evolving grids.Extensive evaluations on eight publicly available smart meter datasets over seven diverse abnormal patterns testing demonstrate the effectiveness of the proposed full framework,achieving average recall and F1 score of more than 0.85.
文摘Research on the range anomaly suppression algorithm in laser radar （ladar） range images is significant in the application and development of ladar. But most of existing algorithms cannot protect the edge and linear target well while suppressing the range anomaly. Aiming at this problem, the differences among the edge, linear target, and range anomaly are analyzed and a novel algo- rithm based on neighborhood pixels detection is proposed. Firstly, the range differences between current pixel and its neighborhood pixels are calculated. Then, the number of neighborhood pixels is detected by the range difference threshold. Finally, whether the current pixel is a range anomaly is distinguished by the neighbor- hood pixel number threshold. Experimental results show that the new algorithm not only has a better range anomaly suppression performance and higher efficiency, but also protects the edge and linear target preferably compared with other algorithms.
基金Projects(61405041,61571145)supported by the National Natural Science Foundation of ChinaProject(ZD201216)supported by the Key Program of Heilongjiang Natural Science Foundation,China+1 种基金Project(RC2013XK009003)supported by Program Excellent Academic Leaders of Harbin,ChinaProject(HEUCF1508)supported by the Fundamental Research Funds for the Central Universities,China
文摘For anomaly detection,anomalies existing in the background will affect the detection performance.Accordingly,a background refinement method based on the local density is proposed to remove the anomalies from thebackground.In this work,the local density is measured by its spectral neighbors through a certain radius which is obtained by calculating the mean median of the distance matrix.Further,a two-step segmentation strategy is designed.The first segmentation step divides the original background into two subsets,a large subset composed by background pixels and a small subset containing both background pixels and anomalies.The second segmentation step employing Otsu method with an aim to obtain a discrimination threshold is conducted on the small subset.Then the pixels whose local densities are lower than the threshold are removed.Finally,to validate the effectiveness of the proposed method,it combines Reed-Xiaoli detector and collaborative-representation-based detector to detect anomalies.Experiments are conducted on two real hyperspectral datasets.Results show that the proposed method achieves better detection performance.
基金Project(1212010071012) supported by Guangdong Pangxidong Mineral Prospect Investigation, ChinaProject(41004051) supported by the National Natural Science Foundation of ChinaProject ([2007]038-01-18) supported by Nationwide Mineral Resource Potential Evaluation Projects of Ministry of Land and Resources, China
文摘Distinguishing geochemical anomalies from background is a basic task in exploratory geochemistry. The derivation of geochemical anomalies from stream sediment geochemical data and the decomposition of these anomalies into their component patterns were described. A set of stream sediment geochemical data was obtained for 1 880 km 2 of the Pangxidong area, which is in the southern part of the recently recognized Qinzhou-Hangzhou joint tectonic belt. This belt crosses southern China and tends to the northwest (NE) direction. The total number of collected samples was 7 236, and the concentrations of Ag, Au, Cu, As, Pb and Zn were measured for each sample. The spatial combination distribution law of geochemical elements and principal component analysis (PCA) were used to construct combination models for the identification of combinations of geochemical anomalies. Spectrum-area (S-A) fractal modeling was used to strengthen weak anomalies and separate them from the background. Composite anomaly modeling was combined with fractal filtering techniques to process and analyze the geochemical data. The raster maps of Au, Ag, Cu, As, Pb and Zn were obtained by the multifractal inverse distance weighted (MIDW) method. PCA was used to combine the Au, Ag, Cu, As, Pb, and Zn concentration values. The S-A fractal method was used to decompose the first component pattern achieved by the PCA. The results show that combination anomalies from a combination of variables coincide with the known mineralization of the study area. Although the combination anomalies cannot reflect local anomalies closely enough, high-anomaly areas indicate good sites for further exploration for unknown deposits. On this basis, anomaly and background separation from combination anomalies using fractal filtering techniques can provide guidance for later work.
基金Project(50374079) supported by the National Natural Science Foundation of China
文摘It is difficult to detect the anomalies whose matching relationship among some data attributes is very different from others’ in a dataset. Aiming at this problem, an approach based on wavelet analysis for detecting and amending anomalous samples was proposed. Taking full advantage of wavelet analysis’ properties of multi-resolution and local analysis, this approach is able to detect and amend anomalous samples effectively. To realize the rapid numeric computation of wavelet translation for a discrete sequence, a modified algorithm based on Newton-Cores formula was also proposed. The experimental result shows that the approach is feasible with good result and good practicality.
基金supported by the National Natural Science Foundation of China(7190121061973310).
文摘Solar arrays are important and indispensable parts of spacecraft and provide energy support for spacecraft to operate in orbit and complete on-orbit missions.When a spacecraft is in orbit,because the solar array is exposed to the harsh space environment,with increasing working time,the performance of its internal electronic components gradually degrade until abnormal damage occurs.This damage makes solar array power generation unable to fully meet the energy demand of a spacecraft.Therefore,timely and accurate detection of solar array anomalies is of great significance for the on-orbit operation and maintenance management of spacecraft.In this paper,we propose an anomaly detection method for spacecraft solar arrays based on the integrated least squares support vector machine(ILS-SVM)model:it selects correlated telemetry data from spacecraft solar arrays to form a training set and extracts n groups of training subsets from this set,then gets n corresponding least squares support vector machine(LS-SVM)submodels by training on these training subsets,respectively;after that,the ILS-SVM model is obtained by integrating these submodels through a weighting operation to increase the prediction accuracy and so on;finally,based on the obtained ILS-SVM model,a parameterfree and unsupervised anomaly determination method is proposed to detect the health status of solar arrays.We use the telemetry data set from a satellite in orbit to carry out experimental verification and find that the proposed method can diagnose solar array anomalies in time and can capture the signs before a solar array anomaly occurs,which reflects the applicability of the method.
文摘Quantitative descriptions of geochemical patterns and providing geochemical anomaly map are important in applied geochemistry. Several statistical methodologies are presented in order to identify and separate geochemical anomalies. The U-statistic method is one of the most important structural methods and is a kind of weighted mean that surrounding points of samples are considered in U value determination. However, it is able to separate the different anomalies based on only one variable. The main aim of the presented study is development of this method in a multivariate mode. For this purpose, U-statistic method should be combined with a multivariate method which devotes a new value to each sample based on several variables. Therefore, at the first step, the optimum p is calculated in p-norm distance and then U-statistic method is applied on p-norm distance values of the samples because p-norm distance is calculated based on several variables. This method is a combination of efficient U-statistic method and p-norm distance and is used for the first time in this research. Results show that p-norm distance of p=2(Euclidean distance) in the case of a fact that Au and As can be considered optimized p-norm distance with the lowest error. The samples indicated by the combination of these methods as anomalous are more regular, less dispersed and more accurate than using just the U-statistic or other nonstructural methods such as Mahalanobis distance. Also it was observed that the combination results are closely associated with the defined Au ore indication within the studied area. Finally, univariate and bivariate geochemical anomaly maps are provided for Au and As, which have been respectively prepared using U-statistic and its combination with Euclidean distance method.
文摘This paper reports statistical results of Seismo-Ionospheric Anomalies(SIAs) of the Total Electron Content(TEC) in the Global Ionosphere Map(GIM) associated with 56 M≥6.0 earthquakes in China during 1998—2012.To detect SIA,a quartile-based(i.e.median-based) process is performed.TEC anomalies for the period of earthquakes without being led by magnetic storms about 10 days are further isolated and examined to confirm the SIP existence.Results show that SIA is the TEC significantly decrease in the afternoon period 2—9 days before the earthquakes in China,which is in a good agreement with the SIA appearing before the 12 May 2008 M 8.0 Wenchuan earthquake.
文摘From Octobet 1998 to January 1999,5 earthquakes ( M s≥5) occurred between Ninglang and Yanyuan counties (27°07′～27°12′N,100°40′～101°00′E area).They were situated in 140km southwest of the Xichang.Among them,the largest one is M s 6 2 on November 19,1998.Based on small seismic data by the seismic remote sensing station of Xichang and the seismological station of Muli,and regional observation data,passing through careful observation and scientific analyses,we had made better forecasts before the earthquakes.That results obvious social benefits.By processing data of precursory earthquakes,such as,original observation data of total geomagnetic intensity from the station of Xichang,pressure capacitance stressometer and quartz horizaontal pendulum tiltmeter from the Xiaomiao station of Xichang,we summarized the sequence characteristics of the series earthquakes.The information about short\|term anomaly of gruond strain,total geomagnetic intensity and ground tilt before the earthquake is emphatically explained.
文摘In recent years,anomaly detection has attracted much attention in industrial production.As traditional anomaly detection methods usually rely on direct comparison of samples,they often ignore the intrinsic relationship between samples,resulting in poor accuracy in recognizing anomalous samples.To address this problem,a knowledge distillation anomaly detection method based on feature reconstruction was proposed in this study.Knowledge distillation was performed after inverting the structure of the teacher-student network to avoid the teacher-student network sharing the same inputs and similar structure.Representability was improved by using feature splicing to unify features at different levels,and the merged features were processed and reconstructed using an improved Transformer.The experimental results show that the proposed method achieves better performance on the MVTec dataset,verifying its effectiveness and feasibility in anomaly detection tasks.This study provides a new idea to improve the accuracy and efficiency of anomaly detection.
基金funded by the National Natural Science Foundation of China, grant number 11975307the National Defense Science and Technology Innovation Special Zone Project, grant number 19-H863-01-ZT-003-003-12。
文摘The anomaly detection of electromagnetic environment situation(EMES) has essential reference value for electromagnetic equipment behavior cognition and battlefield threat assessment.In this paper,we proposed a deep learning-based method for detecting anomalies in EMES to address the problem of relatively low efficiency of electromagnetic environment situation anomaly detection(EMES-AD).Firstly,the convolutional kernel extracts the static features of different regions of the EMES.Secondly,the dynamic features of the region are obtained by using a recurrent neural network(LSTM).Thirdly,the Spatio-temporal features of the region are recovered by using a de-convolutional network and then fused to predict the EMES.The structural similarity algorithm(SSIM) is used to determine whether it is anomalous.We developed the detection framework,de-signed the network parameters,simulated the data sets containing different anomalous types of EMES,and carried out the detection experiments.The experimental results show that the proposed method is effective.
文摘图像异常检测旨在识别并定位图像中的异常区域,针对现有算法中不同层次特征信息利用不充分的问题,提出了基于多层次特征融合网络的图像异常检测算法。通过使用融合了异常先验知识的伪异常数据生成算法,对训练集进行了异常数据扩充,将异常检测任务转化为监督学习任务;构建了多层次特征融合网络,将神经网络中不同层次特征进行融合,丰富了特征中的低层纹理信息和高层语义信息,使得用于异常检测的特征更具区分性;训练时,设计了分数约束损失和一致性约束损失,并结合特征约束损失对整个网络模型进行训练。实验结果表明,MVTec数据集上图像级检测接收机工作特性曲线下面积(area under the receiver operating characteristic, AUROC)平均值为98.7%,像素级定位AUROC平均值为97.9%,每区域重叠率平均值为94.2%,均高于现有的异常检测算法。
