23results about How to "Improve anomaly detection performance" patented technology

The invention discloses an anomaly detection and repair method for non-Gaussian multi-mode sensor data, and belongs to the technical field of unmanned system multi-source autonomous navigation. The method solves the problems that an existing method is poor in performance when facing non-Gaussian data, anomaly detection is difficult to carry out on high-dimensional unstructured data, the robustness of a static anomaly detection model is insufficient, and redundancy compensation is lacked when abnormal data of a sensor is repaired. The method specifically comprises the following steps: step 1, acquiring IMU, GNSS, point cloud and image data of an unmanned system, and then carrying out preprocessing and time synchronization on the acquired data; step 2, respectively extracting features of IMU, GNSS, point cloud and image data at each acquisition moment after time synchronization; step 3, fusing the data features at each acquisition moment; and 4, performing anomaly detection on the data at each acquisition moment according to a fusion result, and further performing data restoration if the data is detected to be abnormal. The method can be applied to data anomaly detection and repair.

The invention discloses a sensor anomaly detection method based on a conditional diffusion probability model and a related system, the system comprises a data acquisition processing unit, a detection model unit and an anomaly detection unit, anomaly is detected by using the conditional probability diffusion model, noise is predicted by adding and removing noise and using a U-Net model, and data is reconstructed. Completing the training of a detection model, and obtaining a trained detection model; and taking the distance between the to-be-detected data and the noise reconstruction data obtained by the trained detection model corresponding to the to-be-detected data as an abnormal score, and if the abnormal score exceeds a threshold value, determining that the data is abnormal. According to the invention, the monitoring accuracy is improved, the false alarm is reduced, the anti-noise capability is enhanced, the noise is effectively identified, and the pertinence is high.

The invention provides a time sequence anomaly detection method based on time-frequency fusion, and belongs to the field of time sequence data anomaly detection. The invention provides a time sequence anomaly detection method based on time-frequency fusion aiming at the problems that potential correlation exists among variables in multivariable time sequence data and frequency domain characteristics are not considered in a traditional method. According to the method, time domain data and frequency domain data in two different modes are innovatively calculated through Spearman correlation coefficients to obtain a time domain similarity matrix and a frequency domain similarity matrix, then fusion is carried out in a point multiplication mode, and potential relations between different variables are mined by using a graph neural network. The time sequence diagram neural network is used for capturing the time dependence and the space relation at the same time, and the features of the time sequence data can be better extracted. The network model is applied to anomaly detection of multivariable time series data, the purpose of improving the accuracy rate can be achieved, and the effectiveness of the method is verified through experimental results.

The invention relates to the technical field of data processing, in particular to a metal alloy current detection resistor detection data processing method and system. The method comprises the following steps: acquiring a plurality of current data and voltage drop data, and obtaining each sub-window of each current data; obtaining the stability degree of each sub-window of each piece of current data, and obtaining the authenticity of each sub-window of each piece of current data according to the stability degree of each sub-window of each piece of current data and the relation between the numerical value change of the current data and the numerical value change of the voltage data; acquiring a reference window of each current data according to the authenticity; according to the method, the fuzzy entropy value of each current data is obtained according to the reference window of each current data, each abnormal data is obtained according to the fuzzy entropy value of each current data, the judgment of the abnormal condition of the metal alloy current detection resistor is completed, the noise influence is avoided, and the accuracy of abnormal detection is improved.

The invention discloses a heat supply pipe network line anomaly judgment method based on temperature and pressure detection, and relates to the technical field of pipeline anomaly detection, and the method comprises the following steps: S1, scanning a two-dimensional code of a to-be-detected point, and reading equipment information of the to-be-detected point; s2, reading real-time temperature data and real-time pressure data measured by a built-in sensor of a point to be detected; s3, analyzing a reasonable temperature difference value between the pipeline branch where the to-be-detected point is located and the heat source by acquiring environmental data of the to-be-detected point; s4, analyzing a reasonable pressure difference value between the pipeline branch where the to-be-detected point is located and the heat source by acquiring terrain data of the to-be-detected point; and S5, according to the detected points in the distribution diagram of the heat supply network, verifying the abnormity judgment result of the current pipeline branch, and after the verification is passed, marking the abnormity information of the pipeline branch. Real-time temperature and pressure data are collected, and dynamic analysis is performed according to a set sampling period, so that the real-time performance and accuracy of a monitoring result are ensured.

The invention discloses a Markov chain-based power dispatching boundary data behavior anomaly detection method and system. The method comprises the following steps: monitoring historical data of each information object of network boundary equipment of a power dispatching system in an IEC104 protocol, and extracting each dispatching service according to a time interval of a remote control instruction data packet of the information object; constructing a state transition matrix based on all historical remote control instructions for each information object based on a Markov chain algorithm, wherein the state transition matrix records all possible transitions among states and probability distribution thereof; when a new scheduling service is received, calculating the probability that the scheduling service belongs to a normal behavior by using the state transition matrix, and if the calculated probability is lower than a preset safety threshold, determining that the scheduling service has an abnormal behavior; and updating the corresponding state transition matrix according to the abnormal behavior detection result. According to the scheme, the anomaly detection capability of the power dispatching control system is remarkably improved.

The invention relates to the technical field of unmanned aerial vehicle detection, in particular to an unmanned aerial vehicle flight data high-precision anomaly detection and data recovery method. Comprising the following steps that S1, a multiple regression model based on TCN-KANs is constructed, the TCN-KANs model comprises an input layer, three residual blocks and a KAN layer, the residual blocks extract features through a convolution layer and residual connection, nonlinear change is carried out on the input features, a potential mode is extracted, final output is generated, and the KAN layer serves as a final output layer of TCN-KANs; s2, minimizing an error between a predicted value and a real value; s3, calculating a residual error according to a prediction result of the model, carrying out smoothing processing on the residual error, and analyzing statistical characteristics of the residual error to determine an abnormal threshold value; and S4, during real-time flight monitoring and operation control of the unmanned aerial vehicle, generating a recovery value for data recovery after abnormality is detected through the model.

The invention relates to the technical field of data anomaly detection, and particularly discloses a multivariate time series anomaly detection method, system and product, and the method comprises the steps: building an anomaly detection framework which comprises a domain attention-based feature extractor, a domain attention-based feature predictor and a multi-domain anomaly discriminator which are connected in sequence; training the domain attention-based feature extractor and the domain attention-based feature predictor by using multivariate time series data of a source domain and a target domain to obtain a multivariate time series anomaly detection model; and performing multivariate time sequence anomaly detection by using the multivariate time sequence anomaly detection model. According to the method, a target domain with limited data and a plurality of source domains with rich data are aligned in distribution, so that cross reuse of data among different devices in a device group is realized; and the model is trained through a joint loss function constrained by domain attention to capture more comprehensive equipment state information.

The invention discloses a network and log intelligent management method. The method comprises the following steps: deploying a network monitoring tool and a log analysis tool in a target network; defining a network monitoring index, monitoring a target network in real time through a network monitoring tool, and obtaining log data through a log monitoring tool; carrying out fusion analysis on the network monitoring information and the log data, and judging whether the target network is abnormal or not according to an analysis result; and defining an alarm mechanism, and giving an alarm when the target network is abnormal. Abnormalities in a system and a network can be quickly found through monitoring, the processing time is shortened, the availability and stability of the system and the network are improved, and configuration optimization can be carried out on network resources through monitoring; potential problems can be identified and prevented through data analysis; the network monitoring information and the log data are fused and analyzed, so that the anomaly detection capability can be further enhanced; and by establishing a sound monitoring and emergency alarm mechanism, the continuity and stability of the service are ensured.

The invention discloses a photovoltaic panel defect detection method and system based on an infrared camera carried by an unmanned aerial vehicle, and the method comprises the steps: collecting an EL image of a photovoltaic panel through the infrared camera carried by the unmanned aerial vehicle, building an image processing algorithm model, recognizing an abnormal region on the photovoltaic panel, and classifying faults; an artificial intelligence technology is utilized to accurately position an abnormal area on the photovoltaic panel, and finally an edge cloud cooperation strategy is adopted to realize lightweight deployment, so that the unmanned nondestructive detection and identification method capable of rapidly and accurately detecting the defects of the photovoltaic panel is provided.

The invention discloses a data transfer method based on a distributed asynchronous architecture, and the data transfer method based on the distributed asynchronous architecture comprises the following steps: S1, building a plurality of computer nodes; s2, when a source node transfers to a target node, monitoring data consistency by using a machine learning model, and detecting an abnormal mode and a potential data error; s3, fragmenting the data needing to be transferred, creating a copy in the source node, starting an asynchronous copying process, and copying the data from the source node to a target node; and S4, after the data transfer is completed, verifying the data in the target node, and ensuring the integrity and consistency of the data again. Through the asynchronous communication and monitoring system, the efficiency and reliability of data transfer are improved, the data consistency is monitored through the machine learning model, the exception detection and error correction capabilities in the data transfer process are enhanced, and the integrity and consistency of the data are ensured.

The invention belongs to the technical field of target detection, and discloses a hyperspectral remote sensing abnormal small target detection method based on multi-aperture spectral information fusion. The method comprises the following steps: dividing a target region into n target sub-regions; collecting hyperspectral data of each target sub-region; converting each piece of hyperspectral data into a spectral data cube, and extracting spatial spectral feature data; extracting difference data from each piece of hyperspectral data; performing anomaly detection according to the spatial spectrum characteristic data and the difference data; according to the invention, through block processing and multi-aperture information fusion of hyperspectral data, the robustness of anomaly detection under spectrum aliasing interference is enhanced; a deep learning technology is introduced, so that higher detection precision and a lower false alarm rate are realized; the LSTM neural network is integrated, so that the dynamic target processing capability is improved; and model parameters are adjusted by adopting a clone selection algorithm, so that the accuracy and robustness of an anomaly detection result are further ensured.

The invention discloses an unsupervised distributed aerial optical cable anomaly detection method, and relates to the field of aerial optical cable state monitoring based on distributed optical fiber sensing, and the method comprises the steps: collecting and demodulating to obtain phase signals returned by an optical fiber under different disturbance conditions, dividing the original phase signals into subsequences through a sliding window, calculating the norm of each sub-sequence, and calculating the norm difference between each sub-sequence and the previous adjacent sub-sequence; calculating a mean value, a minimum value, a maximum value and a standard deviation of each sub-sequence on each dimension to obtain a feature sequence, and inputting the obtained feature sequence into an LSTM-AE model; and calculating a root mean square error RMSE between the input sequence and the reconstructed sequence, and judging the sample condition according to the RMSE value. According to the method, the LSTM-AE is trained only by using the normal sample, so that abnormal event detection based on the unlabeled data set is realized.

Provided are an abnormality detection method and device for a variable flow system of a wind turbine generator. The abnormality detection method comprises: determining a temperature threshold of a target component in the variable flow system based on power of the wind turbine generator; determining whether the temperature threshold of the target component and an operating temperature of the target component satisfy a preset condition, wherein the preset condition comprises a difference between the operating temperature and the temperature threshold being greater than a preset difference; in response to the temperature threshold and the operating temperature satisfying the preset condition, determining whether an operating state of a water-cooling heat dissipation system of the wind turbine generator is abnormal; and in response to the operating state of the water-cooling heat dissipation system being normal, determining that the target component has an abnormal temperature. The abnormality detection method and device for the variable flow system of the wind turbine generator improve the effect of abnormality detection of the wind turbine generator.

The embodiment of the application provides a kind of detection model training method, comprising: obtaining the sample data corresponding to sample graph structure, wherein the initial node attribute of each node and the initial edge attribute of each edge are included in sample data.The initial node attribute is input into the encoding unit of detection model, to obtain the first feature vector output by encoding unit, and the initial edge attribute is input into the encoding unit of detection model, to obtain the second feature vector output by encoding unit.According to the first feature vector and the second feature vector, reconstruct the sample graph structure, and determine the reconstruction error corresponding to the reconstructed sample graph structure.According to the first feature vector and the second feature vector, determine the semantic information corresponding to the sample graph structure, and determine the semantic error corresponding to the sample graph structure according to the semantic information.According to reconstruction error and semantic error, update the model parameter of detection model.The technical scheme of the application can effectively improve the anomaly detection effect of detection model.

The application discloses a kind of space-time perception framework for multivariate time series anomaly detection, belong to abnormal detection technical field, comprising: variable community analysis driven grouping remodeling module, for variable spatial relationship modeling and variable grouping to multivariate time series, and each variable group is respectively reconstructed;Error enhancement module driven by space-time stationarity module, for spatial importance weighting and time stationarity smoothing to reconstruction error, and generate enhanced anomaly score.The application explicitly utilizes the spatial correlation between variables by variable community analysis driven grouping reconstruction, groups strongly correlated variables and independently reconstructs, effectively reduces the noise and interference introduced by weakly correlated variables, and improves the reconstruction quality and model generalization ability.Two modules that capture space-time relationship features work together, significantly improve the accuracy and robustness of anomaly detection.

This invention relates to a method and related equipment for detecting abnormal object behavior. The method includes: acquiring a behavior log stream and mapping the logs to feature vectors including object identifiers; for any object, generating a behavior state sequence based on the feature vectors, constructing reachable links between state points based on the distance relationships between state points in the behavior state sequence, thereby constructing a topological structure of the object's behavior state space, calculating the topological closure of the object's historical state vector set under the topological structure, and using the topological closure as the normal behavior baseline region of the object; wherein, the behavior state space consists of multiple state points obtained by mapping state vectors from the behavior state sequence; acquiring the target state vector corresponding to the behavior of the object to be detected, determining whether the target state vector belongs to the normal behavior baseline region of the corresponding object, and if not, determining it as abnormal behavior. This method at least partially solves the problems of high storage overhead and poor anomaly detection effect in related technologies.

One or more embodiments of the invention provide an interface abnormity detection method and device. The interface abnormity detection method comprises the steps that for an interaction behavior implemented by a user in a graphical user interface (GUI), a behavior log of the interaction behavior and an interaction video corresponding to the interaction behavior are generated, and the interaction video is used for recording the process that the GUI updates interface content in response to the interaction behavior; and performing data fusion processing on the behavior log and the interactive video to generate behavior sequence data of the user in the GUI, and detecting interface abnormity of the GUI based on the behavior sequence data.

The application discloses an industrial image anomaly detection method and system based on feature-level abnormal synthesis and reconstruction, and the method is as follows: S1, acquiring an image, and extracting multi-scale local features through a pre-trained network; S2, adjusting the features of different scales to a unified size, connecting the features of the unified size, and obtaining a feature map; S3, transferring the feature map of S2 to a target domain through feature adaptation, and obtaining a new feature map; S4, adding noise to the feature map obtained in S3, and obtaining a feature map simulating an anomaly; S5, inputting the feature map obtained in S4 into a feature reconstruction network, and reconstructing the features into features without anomalies; S6, calculating a loss through the difference between the feature map before adding noise and the reconstructed feature map, training the feature reconstruction network according to the loss value; and S7, reconstructing the features by using the trained network, and obtaining a segmentation map according to the difference before and after the feature reconstruction. The application solves the problems of high calculation cost, unreal synthesized anomaly and reconstructed anomaly area of the existing anomaly detection method.

The present application relates to the technical field of data transmission, and discloses a method and system for detecting and eliminating abnormal data under complex working conditions, the method comprising: receiving data streams of multiple sensors in a preset complex working condition operation process; extracting features from the data stream of each sensor in the multiple sensors according to multiple preset time points; finding the feature vectors whose number of feature vectors in all neighborhoods reaches the minimum number of points as core data points according to the preset neighborhood radius and the minimum number of points, clustering the multiple feature vectors to form multiple feature clusters and noise data points outside the multiple feature clusters; calculating the outlier measure of the noise data point; calculating the probability that the noise data point is an abnormal data point; when the probability that the noise data point is an abnormal data point is higher than a preset threshold, according to the time point corresponding to the abnormal data point, eliminating data from the data streams of the multiple sensors. The present invention effectively improves the data quality and enhances the abnormality detection capability.

The invention relates to the technical field of automatic detection, and particularly discloses a rotary tower type spray visual automatic detection system of a perfume sprayer, which comprises a visual detection module for capturing an image in a spray process in real time and analyzing a spray area in the image to obtain spray abnormal data; the data processing module is used for calculating an evaluation coefficient based on the abnormal spraying data; the spraying evaluation module is used for evaluating the spraying process based on the evaluation coefficient, judging whether the spraying process is abnormal or not and generating an evaluation signal; the abnormal reason analysis module is used for monitoring visual detection equipment based on the spraying abnormal early warning signal, analyzing whether the generation of the spraying abnormal early warning signal is caused by the visual detection equipment or not, and generating an equipment analysis signal; according to the invention, the abnormity detection capability of the system is improved, the root of the spraying abnormity can be more accurately judged, and a powerful guarantee is provided for stable operation of the perfume sprayer, so that the system can more quickly position the problem when facing the spraying abnormity.

A method for anomaly detection in nuclear power DCS logs based on weighted reconstruction and state transition probabilities includes: acquiring DCS controller log data; preprocessing the log data, including log parsing, log template extraction, and log annotation, to obtain structured log data, log templates, and anomaly tags; constructing log sequences, including device grouping and time window segmentation, converting DCS logs into structured sequences, and dividing them into training, testing, and validation sets; constructing a sparse-based hybrid weighting mechanism, and using this mechanism to train a weighted LSTM autoencoder on the training set; constructing a Markov transition matrix to comprehensively extract deep temporal reconstruction features and local state transition patterns of the log sequences; using the validation set for dynamic threshold optimization to determine the optimal anomaly judgment boundary; and completing anomaly detection and performance evaluation on the testing set based on the optimized joint model. This method can effectively improve the accuracy and reliability of anomaly detection in nuclear power DCS systems.

The present invention discloses a dual prevention system for enterprise safety production, relates to the technical field of dual prevention, and comprises a control center, wherein the control center is connected with a database, a standard behavior model generation module, a real-time behavior data acquisition module, a real-time behavior data processing module and a risk processing module; the database is used to store risk management information; the standard behavior model generation module is used to generate standard behavior data corresponding to employees and equipment respectively according to the risk management information, and generate a standard behavior model according to the standard behavior data; the real-time behavior data acquisition module is used to obtain the real-time behavior data of employees and equipment respectively in real time; the real-time behavior data processing module is used to determine whether there is abnormal behavior in the real-time behavior data through the standard behavior model, and if there is abnormal behavior, generate corresponding risk warning according to the abnormal behavior; the risk processing module is used to generate corresponding risk resolution strategy according to the risk warning.