High-precision anomaly sensing operation and maintenance method and system based on unmanned aerial vehicle inspection

Abstract

The invention discloses a high-precision anomaly sensing operation and maintenance method and system based on unmanned aerial vehicle inspection, and relates to the technical field of operation and maintenance management, and the method comprises the steps: setting a plurality of inspection targets according to an inspection target region, and sequentially mining features and determining an inspection index set; the target unmanned aerial vehicle takes off according to a preset plan through the take-off, landing and charging integrated platform, monitors the inspection target based on the inspection index set, and carries the high-precision sensor group to obtain inspection data. Performing abnormal value identification and preprocessing on the data to generate a detection data set; and training a multi-source heterogeneous anomaly perception algorithm model library based on the inspection target, calling the model library to perform matching analysis on the detection data set, and generating an anomaly perception report. The report is transmitted to an operation and maintenance center, linkage operation and maintenance strategy parameters are obtained through analysis, and regional abnormal operation and maintenance management and control are executed. Therefore, the technical effects of reducing the detection difficulty and improving the anomaly perception timeliness and precision are achieved.

Description

Technical Field

[0001] The present invention relates to the field of operation and maintenance management technology, and in particular to a high-precision abnormality perception operation and maintenance method and system based on unmanned aerial vehicle inspection. Background Art

[0002] Regional anomaly detection plays a key role in security prevention. In order to meet the needs of efficient, safe and precise operation and maintenance, drone inspections have been widely used, which can significantly improve inspection efficiency, reduce operation and maintenance costs, and reduce the safety risks of manual inspections.

[0003] Existing drone inspections mostly use high-precision cameras, infrared thermal imagers, lidar and other sensors to collect large amounts of images, videos and environmental data in real time, and detect and identify anomalies from these data. Faced with massive multi-dimensional data, there are technical problems such as high detection difficulty, low timeliness and impact on perception accuracy. Summary of the invention

[0004] The present invention provides a high-precision anomaly perception operation and maintenance method and system based on drone inspection to solve the technical problems in the prior art of high detection difficulty, low timeliness, and impact on perception accuracy, and achieve the technical effect of reducing detection difficulty and improving the timeliness and accuracy of anomaly perception.

[0005] In a first aspect, the present invention provides a high-precision abnormality perception operation and maintenance method based on drone inspection, wherein the method comprises:

[0006] According to the inspection target area, multiple inspection targets are set, and feature mining is performed on the multiple inspection targets in turn to determine multiple target inspection indicator sets.

[0007] The target UAV takes off at a scheduled time according to a preset inspection plan through the integrated take-off, landing and charging platform, and performs inspection and monitoring on the multiple inspection targets based on the multiple target inspection indicator sets, wherein the target UAV is equipped with a high-precision sensor group.

[0008] A plurality of target inspection data sets are acquired by using the high-precision sensor group for inspection, and outlier identification and data preprocessing are performed on the plurality of target inspection data sets according to the plurality of target inspection indicator sets to generate a plurality of target detection data sets.

[0009] A multi-source heterogeneous anomaly perception algorithm model library is constructed based on the multiple inspection target training, and the multi-source heterogeneous anomaly perception algorithm model library is called to perform matching analysis and processing on the multiple target detection data sets respectively to obtain multiple target anomaly perception reports.

[0010] The multiple target abnormality perception reports are transmitted to the operation and maintenance center for strategy analysis and optimization, to obtain regional linkage operation and maintenance strategy parameters, and based on the regional linkage operation and maintenance strategy parameters, abnormal operation and maintenance control is performed on the inspection target area.

[0011] In a second aspect, the present invention also provides a high-precision abnormality perception operation and maintenance system based on drone inspection, wherein the system comprises:

[0012] The target indicator extraction module is used to set multiple inspection targets according to the inspection target area, and perform feature mining on the multiple inspection targets in turn to determine multiple target inspection indicator sets.

[0013] An inspection plan execution module is used for the target UAV to take off on time according to a preset inspection plan through an integrated take-off, landing and charging platform, and to perform inspection and monitoring on the multiple inspection targets based on the multiple target inspection indicator sets, wherein the target UAV is equipped with a high-precision sensor group.

[0014] The inspection data acquisition module is used to obtain multiple target inspection data sets by using the high-precision sensor group inspection, perform outlier identification and data preprocessing on the multiple target inspection data sets according to the multiple target inspection indicator sets, and generate multiple target detection data sets.

[0015] An anomaly perception module is used to build a multi-source heterogeneous anomaly perception algorithm model library based on the multiple inspection target training, call the multi-source heterogeneous anomaly perception algorithm model library to perform matching analysis and processing on the multiple target detection data sets respectively, and obtain multiple target anomaly perception reports.

[0016] A decision management module is used to transmit the multiple target abnormality perception reports to the operation and maintenance center for strategy analysis and optimization, obtain regional linkage operation and maintenance strategy parameters, and perform abnormal operation and maintenance control on the inspection target area based on the regional linkage operation and maintenance strategy parameters.

[0017] The present invention discloses a high-precision abnormality perception operation and maintenance method and system based on unmanned aerial vehicle inspection, comprising: setting a plurality of inspection targets according to an inspection target area, and performing feature mining on these inspection targets in turn to determine a plurality of target inspection index sets; the target unmanned aerial vehicle takes off at a scheduled time according to a preset inspection plan through an integrated take-off, landing and charging platform, and performs inspection and monitoring on these inspection targets based on a plurality of target inspection index sets, wherein the target unmanned aerial vehicle is equipped with a high-precision sensor group; collecting inspection data by using the high-precision sensor group to obtain a plurality of target inspection data sets, performing abnormal value identification and data preprocessing on the data sets according to the inspection index set, and generating a plurality of target detection data sets; constructing a multi-source heterogeneous abnormality perception algorithm model library based on inspection target training, calling the model library to perform matching analysis and processing on a plurality of target detection data sets, and generating a plurality of target abnormality perception reports; transmitting the plurality of target abnormality perception reports to an operation and maintenance center, performing strategy analysis and optimization, obtaining regional linkage operation and maintenance strategy parameters, and implementing abnormal operation and maintenance control on the inspection target area according to the parameters. The high-precision anomaly perception operation and maintenance method and system based on drone inspection disclosed in the present invention solve the technical problems of high detection difficulty, low timeliness and impact on perception accuracy, and achieve the technical effect of reducing detection difficulty and improving the timeliness and accuracy of anomaly perception. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a flow chart of the high-precision abnormality perception operation and maintenance method based on drone inspection of the present invention;

[0019] Figure 2 It is a structural schematic diagram of the high-precision abnormality perception operation and maintenance system based on drone inspection of the present invention.

[0020] Explanation of the reference numerals: target index extraction module 11, inspection plan execution module 12, inspection data collection module 13, abnormality perception module 14, decision management module 15. DETAILED DESCRIPTION

[0021] The technical solution provided in the embodiments of the present invention is to solve the technical problems of the prior art, such as high detection difficulty, low timeliness, and impact on perception accuracy. The overall idea adopted is as follows:

[0022] Firstly, according to the inspection target area, multiple inspection targets are set, and the features of these inspection targets are mined in turn to determine multiple target inspection indicator sets; then, the target UAV takes off regularly according to the preset inspection plan through the integrated take-off, landing and charging platform, and performs inspection and monitoring on the inspection target based on multiple target inspection indicator sets, and the UAV is equipped with a high-precision sensor group; then, the high-precision sensor group is used to inspect and obtain multiple target inspection data sets, and outlier identification and preprocessing are performed on these data according to the target inspection indicator set to generate multiple target detection data sets; then, based on the inspection target, a multi-source heterogeneous anomaly perception algorithm model library is trained and constructed, and the model library is called to match and analyze multiple target detection data sets to obtain multiple target anomaly perception reports; then, these reports are transmitted to the operation and maintenance center for strategy analysis and optimization to obtain regional linkage operation and maintenance strategy parameters; finally, based on the regional linkage operation and maintenance strategy parameters, abnormal operation and maintenance control is performed on the inspection target area.

[0023] The above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods of the specification to better understand the above technical solution. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments of the present invention. It should be understood that the present invention is not limited to the example embodiments used only to explain the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. In addition, it should be noted that, for the convenience of description, only the parts related to the present invention are shown in the drawings, rather than all of them.

[0024] Embodiment 1

[0025] Figure 1 The figure is a flow chart of a high-precision abnormality perception operation and maintenance method based on drone inspection of the present invention, wherein the method comprises:

[0026] According to the inspection target area, multiple inspection targets are set, and feature mining is performed on the multiple inspection targets in turn to determine multiple target inspection indicator sets.

[0027] Specifically, first, visit the inspection target area and locate multiple inspection targets, where the inspection target refers to multiple objects to be inspected in a specific place or range that need to be monitored and inspected, including nuclear power sources, equipment sources, fire risk sources, etc., distributed in different locations in the inspection area. The setting and selection of inspection targets are based on the abnormal perception target setting of the inspection target area.

[0028] Specifically, the target inspection indicator set is a set of monitoring and inspection parameters generated for each inspection target, which is used to monitor and evaluate the status of the inspection target in real time, thereby ensuring the safety and stability of its operation. In other words, the target inspection indicator set includes all indicator dimensions that can reflect the status of the inspection target. Multiple inspection targets correspond to multiple target inspection indicator sets, and the target inspection indicator sets of inspection targets of different types, different attributes, and different scale levels are not the same. For example, it includes the power consumption, operating voltage, current, and temperature of the equipment source; the temperature and smoke concentration of the fire source, etc.

[0029] In some embodiments, determining multiple target inspection indicator sets includes:

[0030] Perform characteristic analysis on the multiple inspection targets in sequence to obtain multiple target basic characteristic data sets, wherein the multiple target basic characteristic data sets include location, model specification, application function and historical maintenance record; perform risk impact assessment on the multiple inspection targets based on the multiple target basic characteristic data sets to obtain multiple target risk level information; determine multiple target feature mining depths according to the multiple target risk level information;

[0031] Based on the multiple target feature mining depths, the multiple target basic characteristic data sets are respectively mined for associated indicators to obtain the multiple target inspection indicator sets.

[0032] Specifically, first, analyze the basic information and operating status of each inspection target, identify the geographical or system location information of each inspection target (such as GPS coordinates, equipment number); analyze the model specifications of the equipment or system, and obtain its design parameters, such as voltage, capacity, size, etc.; determine the core application functions of the inspection target. For example, for nuclear power equipment, the functions include power supply and stability assurance; for fire detectors, the function is to monitor fire risks; retrieve the maintenance history data of the target, including past fault repairs, maintenance frequency, etc., and then form the basic characteristic data set of the system. The information in this basic characteristic data set helps to judge the health status and failure possibility of the equipment.

[0033] Specifically, based on the basic characteristic data of the target, its operating status and potential risks are evaluated, and the risk level of the target is determined, which directly affects the priority and depth of the inspection. For example, based on location, function, and historical maintenance records, it is evaluated whether the equipment is currently in a risky state (maintenance interval is too long, equipment model is too old, or load is too large), and each inspection target is assigned to different risk levels according to the risk level. Such as high risk, medium risk, and low risk.

[0034] Specifically, the feature mining depth determines the correlation range and the number of indicators to be mined when setting inspection indicators for each inspection target. High-risk equipment requires in-depth mining of more related features and indicators to ensure that multiple related indicators can be covered; while low-risk equipment can only focus on core indicators and basic health status.

[0035] Furthermore, according to the determined feature mining depth, related indicators are mined from the target's basic characteristic data set, and the final inspection indicator set is generated by analyzing the health status, performance, safety compliance, etc. of the equipment. Preferably, related indicator mining is performed on multiple target basic characteristic data sets through methods such as association analysis and principal component analysis, so as to obtain multiple target inspection indicator sets that meet the feature mining depth and have the highest degree of correlation.

[0036] The above method steps determine the appropriate feature mining depth through characteristic analysis and risk assessment of the inspection target, ensuring the comprehensiveness and pertinence of the inspection indicator set. At the same time, according to different risk levels, the mining depth and the number of inspection indicators are dynamically adjusted to improve the inspection efficiency while ensuring that the equipment with higher risks is fully monitored, which helps to ensure that the inspection process can promptly discover potential problems and take measures.

[0037] In some implementations, obtaining the multiple target inspection indicator sets includes:

[0038] Performing associated indicator mining on the multiple target basic characteristic data sets respectively to obtain multiple target associated indicator sets, and clustering and labeling the multiple target associated indicator sets to obtain multiple target classification indicator cluster sets; performing correlation analysis on the multiple target classification indicator cluster sets based on the multiple target basic characteristic data sets to obtain multiple indicator cluster correlation coefficient sets; performing primary indicator screening on the multiple target classification indicator cluster sets according to the multiple target feature mining depths and the multiple indicator cluster correlation coefficient sets to obtain multiple target feature indicator cluster sets; performing correlation analysis and secondary indicator screening on each feature indicator in the multiple target feature indicator cluster sets based on the multiple target feature mining depths to obtain the multiple target inspection indicator sets.

[0039] Specifically, first, we mine the target's basic characteristic data set to extract the indicator sets related to the target's operation, performance, safety, etc., and output them as multiple target-related indicator sets. These indicators are important data sources for target monitoring.

[0040] Specifically, each inspection target includes multiple types of related indicators, such as physical parameters (such as temperature, pressure, vibration), performance parameters (such as current, voltage, throughput), environmental parameters (such as temperature, humidity, light), etc. Therefore, the distribution performs clustering division on the obtained multiple target-related indicator sets within the set, and obtains multiple target classification indicator cluster sets, so as to organize and manage indicators of different categories more clearly.

[0041] Further, a suitable statistical method is selected to calculate the correlation coefficients between the indicator clusters, such as the Pearson correlation coefficient, the Spearman correlation coefficient, and the Kendall correlation coefficient. Exemplarily, first, the data in each pair of indicator clusters are paired. Then, the selected correlation calculation method is used to calculate the correlation coefficients between each pair of indicator clusters, and the calculation results are recorded and stored to form a set of correlation coefficients of multiple indicator clusters.

[0042] Specifically, first, determine the screening range according to the feature mining depth, and set different indicator screening ranges for different inspection targets according to their risk levels and feature mining depths. The screening range for high-risk targets is wider, while fewer core indicators are screened for low-risk targets. Then. For each inspection target, sort the corresponding indicator clusters from high to low according to the correlation coefficient, and screen the indicator cluster with the highest correlation according to the feature mining depth. Each inspection target eventually forms a feature indicator cluster set. For example, if the feature mining depth of a high-risk inspection target is 3, the indicator clusters of the top 3 categories with the highest correlation coefficients are selected from the serialized target classification indicator cluster set as the first-level indicator screening results.

[0043] Furthermore, according to the feature mining depth and the correlation coefficient, the multiple indicators in the target feature indicator cluster set are screened at the second level. In other words, the indicator items with the highest correlation with the anomaly are selected from the multiple indicator clusters of the first-level indicator screening results, thereby obtaining multiple target inspection indicator sets.

[0044] Through the above-mentioned method steps, multi-level correlation analysis and screening are carried out to generate a set of inspection indicators covering the target health status, performance and environmental impact, ensuring that the inspection process is accurate and efficient, comprehensively reflecting the operating status of the inspection target while avoiding data redundancy, and helping to reduce the difficulty of inspection collection and data analysis.

[0045] The target UAV takes off at a scheduled time according to a preset inspection plan through the integrated take-off, landing and charging platform, and performs inspection and monitoring on the multiple inspection targets based on the multiple target inspection indicator sets, wherein the target UAV is equipped with a high-precision sensor group.

[0046] Specifically, the integrated take-off, landing and charging platform is used for take-off, landing and charging of drones. The target drone is equipped with a high-precision sensor group for inspection and monitoring. Among them, the high-precision sensor group includes a high-precision camera, infrared sensor, laser radar, gas sensor, etc. The specific configuration of the high-precision sensor group is determined according to the monitoring target of the target drone.

[0047] Specifically, the preset inspection plan includes information such as inspection time, inspection route and inspection target. The drone flies along the preset inspection route and monitors the inspection target based on multiple target inspection indicator sets through the high-precision sensor group it carries.

[0048] Specifically, after completing the inspection mission, the target UAV returns to the integrated take-off and landing charging platform according to the preset return route, and automatically lands on the integrated take-off and landing charging platform. The platform charges the UAV in preparation for the next inspection mission.

[0049] A plurality of target inspection data sets are acquired by using the high-precision sensor group for inspection, and outlier identification and data preprocessing are performed on the plurality of target inspection data sets according to the plurality of target inspection indicator sets to generate a plurality of target detection data sets.

[0050] Specifically, a high-precision sensor group configured on the target drone is used to inspect and collect multiple target inspection data sets of multiple inspection targets, wherein the data types included in the multiple target inspection data sets correspond to the multiple target inspection indicator sets. For example, the target inspection data sets may include temperature data, image data, location data, ambient air data, etc.

[0051] Specifically, the data collection frequency of the sensors in the high-precision sensor group is set according to the characteristics and requirements of the inspection target. For high-risk or rapidly changing targets, the collection frequency should be set higher; for stable running targets, the collection frequency can be appropriately reduced.

[0052] In some embodiments, generating multiple target detection datasets includes:

[0053] According to the multiple target inspection indicator sets, multiple target data anomaly identification factors are set, and the multiple target data anomaly identification factors include data consistency, noise data and range rationality; based on the multiple target data anomaly identification factors, outlier identification is performed on the multiple target inspection data sets to obtain multiple target anomaly data sets; according to the multiple target data anomaly identification factors, multiple target anomaly data preprocessing programs are mapped to obtain multiple target anomaly data sets, and data cleaning is performed on the multiple target anomaly data sets based on the multiple target anomaly data preprocessing programs to obtain multiple available target data sets; feedback correction is performed on the multiple available target data sets based on the sensor empirical error factor to generate the multiple target detection data sets.

[0054] Specifically, data anomaly identification factors are used to identify and determine possible anomalies in the inspection data set, including data consistency (whether the data is consistent), noise data (whether the data is affected by noise), and range rationality (whether the data deviates from the normal range).

[0055] Specifically, compare the detection results between different sensors to determine whether there are obvious inconsistencies in the data. For example, through statistical indicators such as variance and standard deviation, calculate and determine whether the data of multiple temperature sensors are significantly different. By setting a noise recognition algorithm, identify short-term extreme values ​​or unreasonable fluctuations in the inspection data set, and filter out noise data. Set a reasonable numerical range for each type of data and define the rationality of the range. If a certain data exceeds the reasonable numerical range, it is judged as abnormal. Then, organize and save the identified abnormal data as an abnormal data set for subsequent data processing and analysis.

[0056] Furthermore, according to different abnormal identification factors, corresponding pre-processing procedures are automatically mapped. Exemplarily, pre-processing for noise data may include outlier removal, wavelet denoising, etc., while pre-processing for data consistency includes data smoothing and correction based on interpolation methods. Based on the characteristics of abnormal data sets, generating appropriate data cleaning procedures helps to ensure that abnormal data can be restored to a reasonable state or eliminated after processing.

[0057] Specifically, the sensor may generate cumulative errors during long-term use, and then the data is further corrected based on the empirical error factors accumulated from historical data. For example, first, based on the long-term operating history data of the sensor, the deviation law of the sensor is calculated (for example, the temperature sensor may measure higher values ​​in a high temperature environment). Then, the empirical error factor is used to adjust each data in the available target data set by a certain proportion to achieve feedback correction and further improve the accuracy of the data.

[0058] The above method steps identify anomalies through multiple identification factors (such as consistency, noise, and range rationality), ensuring that outliers in the data can be found quickly and accurately, and abnormal data can be removed or corrected in a targeted manner, ensuring the quality of the cleaned data. The feedback correction of the sensor's empirical error factor further improves the accuracy of the data, ensuring that the final detection data set can be used for accurate decision-making and analysis.

[0059] A multi-source heterogeneous anomaly perception algorithm model library is constructed based on the multiple inspection target training, and the multi-source heterogeneous anomaly perception algorithm model library is called to perform matching analysis and processing on the multiple target detection data sets respectively to obtain multiple target anomaly perception reports.

[0060] In some embodiments, a multi-source heterogeneous anomaly perception algorithm model library is constructed, including:

[0061] Based on the multiple target inspection indicator sets, multi-source heterogeneous data collection is performed to obtain multiple target multi-source heterogeneous anomaly data sets; anomaly perception feature extraction is performed on the multiple target multi-source heterogeneous anomaly data sets to obtain multiple target anomaly perception data feature sets; feature algorithm selection and anomaly model training are performed on the multiple target multi-source heterogeneous indicator data sets according to the multiple target anomaly perception data feature sets to obtain multiple target anomaly recognition model sets; based on the multiple inspection targets, the multiple target anomaly recognition model sets are integrated to construct the multi-source heterogeneous anomaly perception algorithm model library.

[0062] Specifically, first, multiple target inspection indicator sets are used as index targets of sample data, historical inspection data are parsed, and multiple target multi-source heterogeneous anomaly data sets are obtained, where the historical inspection data includes inspection logs of multiple inspection targets, anomaly lists and corresponding anomaly handling records.

[0063] Specifically, the anomaly perception feature is extracted from the multi-source heterogeneous anomaly data set and is highly correlated with the abnormal state of the target device or system. By extracting anomaly perception features from multiple target multi-source heterogeneous anomaly data sets, the data part of the original sample data that contributes to anomaly monitoring is retained, and data simplification of multiple target multi-source heterogeneous anomaly data sets is achieved. The extracted anomaly perception data feature set has high data quality, which helps to improve the efficiency of the subsequent training process and avoid overfitting.

[0064] Specifically, according to the feature sets of different inspection targets, select a suitable feature extraction algorithm to extract the key features that best reflect the abnormal state. For example, use convolutional neural network (CNN) to process image data, use LSTM to process time series data, and use PointNet to process 3D point cloud data. Optionally, before feature extraction, perform preprocessing such as standardization and normalization on the data to ensure that data from different sources can be processed under the same framework.

[0065] Specifically, the extracted abnormal perception feature set is used to train the abnormal recognition model. During the training process, the model will learn according to the abnormal features of different inspection targets and acquire the ability to recognize abnormalities of different inspection targets. In other words, by learning the mapping relationship between multiple features in the abnormal perception feature set and the corresponding abnormal situations, the model acquires the ability to judge abnormal situations. For example, the model can identify the type, level, range, etc. of the abnormality.

[0066] Furthermore, according to the target types of multiple inspection targets, the anomaly recognition models of different inspection targets are integrated to form a unified model library. Each model in the multi-source heterogeneous anomaly perception algorithm model library provides specific anomaly recognition functions according to the different inspection targets and data sources. By building an anomaly perception algorithm model library that can process multi-source heterogeneous data and adapt to different inspection targets, it is helpful to quickly call the target anomaly recognition model when processing different inspection tasks or inspection targets.

[0067] In some implementations, obtaining multiple target multi-source heterogeneous anomaly data sets includes:

[0068] Mining and acquiring a regional inspection anomaly database, performing first-order correlation collection in the regional inspection anomaly database based on the multiple target inspection indicator sets, and obtaining multiple first-order correlation indicator data sets; taking the multiple first-order correlation indicator data sets as a benchmark, performing second-order correlation collection in the regional inspection anomaly database, and obtaining multiple second-order correlation indicator data sets; until traversing and collecting N-1 order indicator data, where 2≤N≤4, based on the multiple first-order correlation indicator data sets, the multiple second-order correlation indicator data sets, and until the N-1 order indicator data, combining to obtain the multiple target multi-source heterogeneous anomaly data sets.

[0069] Specifically, first, the historical inspection data in the area, including normal data and abnormal data, is collected and sorted to build a regional inspection abnormality database. Then, in the regional inspection abnormality database, first-order correlation collection is performed based on the target inspection indicator set, and correlation analysis, association rule mining and other technologies are used to identify abnormal data directly related to the target inspection indicator, and obtain multiple first-order correlation indicator data sets directly related to the target indicator. The first-order correlation indicator data set is a collection of abnormal indication indicators that are strongly or indirectly related to the inspection target.

[0070] Specifically, based on multiple first-order correlation indicator data sets, second-order correlation collection is performed in the regional inspection anomaly database to obtain multiple second-order correlation indicator data sets related to the first-order correlation indicator data sets. The second-order correlation indicator data set is a collection of anomaly indicator indicators that are weakly or indirectly related to the inspection target.

[0071] Furthermore, the above process is repeated, each time taking the previous first-order correlation indicator data set as the benchmark, the next-order correlation collection is performed, and the scope of the correlation data set is gradually expanded for correlation collection until the N-1 order, and multiple first-order correlation indicator data sets, multiple second-order correlation indicator data sets, and N-1 order indicator data sets are combined to form multiple target multi-source heterogeneous anomaly data sets. Where N is greater than or equal to 2 and less than or equal to 4, that is, the correlation collection is performed at least four times and at most twice. By gradually expanding the scope of the correlation data set and increasing the amount of data collection, more data support is provided for model training, which helps to improve the accuracy and robustness of the model.

[0072] The multiple target abnormality perception reports are transmitted to the operation and maintenance center for strategy analysis and optimization, to obtain regional linkage operation and maintenance strategy parameters, and based on the regional linkage operation and maintenance strategy parameters, abnormal operation and maintenance control is performed on the inspection target area.

[0073] Specifically, multiple target abnormality perception reports generated during the inspection process are transmitted to the operation and maintenance center in real time, and each abnormality perception report contains abnormal characteristic data of the target, such as temperature abnormality, humidity abnormality, image abnormality, etc. Among them, the operation and maintenance center is used to perform operation and maintenance management of the inspection target area, and the operation and maintenance center contains all abnormality handling plans and operation and maintenance strategies for multiple inspection targets in the inspection target area.

[0074] In some embodiments, obtaining regional linkage operation and maintenance strategy parameters includes:

[0075] Based on the abnormal perception reports of the multiple targets, feature matching is performed with the abnormal operation and maintenance strategy library respectively to obtain abnormal operation and maintenance strategies applicable to the multiple targets; according to the abnormal operation and maintenance strategies applicable to the multiple targets, multiple abnormal operation and maintenance strategy spaces are constructed, and parameter analysis is performed on the abnormal perception reports of the multiple targets according to the multiple abnormal operation and maintenance strategy spaces to obtain multiple target operation and maintenance strategy parameter thresholds; global parameter optimization is performed within the multiple target operation and maintenance strategy parameter thresholds according to the preset search step length to obtain multiple target operation and maintenance strategy parameters; the multiple target operation and maintenance strategy parameters are linked and integrated according to the drone inspection order to obtain the regional linkage operation and maintenance strategy parameters.

[0076] Specifically, the operation and maintenance center is equipped with an abnormal operation and maintenance strategy library, which contains operation and maintenance strategies corresponding to various abnormal situations. Furthermore, based on multiple target abnormal perception reports, feature matching is performed with the abnormal operation and maintenance strategy library respectively, and abnormal operation and maintenance strategies applicable to multiple targets can be obtained. Exemplarily, based on the rule matching algorithm, the abnormal features in the perception report are matched with the feature templates in the strategy library to identify the most applicable operation and maintenance strategy, such as the top K operation and maintenance strategies with the highest activation frequency. Preferably, K is determined based on the risk level of the inspection target.

[0077] Specifically, according to the abnormal operation and maintenance strategies applicable to multiple targets, a strategy space is generated for each target abnormal perception report, and multiple abnormal operation and maintenance strategy spaces are generated, wherein the abnormal operation and maintenance strategy space contains all commonly used operation and maintenance strategies and their parameter ranges. Then, the parameter range of each operation and maintenance strategy is parsed, and the upper and lower limits of each parameter are determined to form a parameter threshold set. This parameter threshold set quantitatively characterizes the applicable scope of multiple commonly used operation and maintenance strategies, which helps to accurately select appropriate operation and maintenance strategies.

[0078] Furthermore, an optimization algorithm (such as genetic algorithm, particle swarm optimization, etc.) is used to perform a global search within the above-mentioned multiple target operation and maintenance strategy parameter thresholds according to a preset search step size to find the optimal operation and maintenance strategy parameter combination. Preferably, the preset search step size is an adaptive step size.

[0079] Finally, based on the inspection sequence and target location, the operation and maintenance strategy parameters of each target are linked and integrated to form an overall regional linkage operation and maintenance strategy, which is output as regional linkage operation and maintenance strategy parameters. The regional linkage operation and maintenance strategy parameters include the execution order, path and multiple operation and maintenance strategies of the operation and maintenance tasks. The linkage integration based on the inspection sequence and target location enables the execution of the operation and maintenance tasks to be carried out on the basis of a reliable inspection route, further ensuring the reliability of the operation and maintenance tasks.

[0080] The above method steps obtain the optimal multiple operation and maintenance strategy parameters through strategy analysis and global optimization to improve the operation and maintenance efficiency. Feature matching is performed based on abnormal perception reports and the operation and maintenance strategy library to ensure the accuracy of the operation and maintenance strategy. Through linkage integration, the scheduling of drones and other operation and maintenance equipment is optimized, resource utilization is improved, and the reliability of inspection tasks is ensured.

[0081] In summary, the high-precision abnormality perception operation and maintenance method based on drone inspection provided by the present invention has the following technical effects:

[0082] According to the inspection target area, multiple inspection targets are set, and the features of these inspection targets are mined in turn to determine multiple target inspection indicator sets; the target UAV takes off regularly according to the preset inspection plan through the integrated take-off and landing charging platform, and inspects and monitors these inspection targets based on multiple target inspection indicator sets, wherein the target UAV is equipped with a high-precision sensor group; the high-precision sensor group is used to collect inspection data, obtain multiple target inspection data sets, and perform outlier identification and data preprocessing on the data sets according to the inspection indicator set to generate multiple target detection data sets; based on the inspection target training, a multi-source heterogeneous anomaly perception algorithm model library is constructed, and the model library is called to match and analyze multiple target detection data sets to generate multiple target anomaly perception reports; the multiple target anomaly perception reports are transmitted to the operation and maintenance center for strategy analysis and optimization, and the regional linkage operation and maintenance strategy parameters are obtained, and the abnormal operation and maintenance control of the inspection target area is implemented according to the parameters. Thus, the technical effect of reducing the difficulty of detection and improving the timeliness and accuracy of anomaly perception is achieved.

[0083] Embodiment 2

[0084] Figure 2 This is a schematic diagram of the structure of the high-precision abnormality perception operation and maintenance system based on drone inspection of the present invention. For example, Figure 1 The flowchart of the high-precision abnormality perception operation and maintenance method based on drone inspection in the present invention can be shown as follows: Figure 2 The structure shown is implemented.

[0085] Based on the same concept as the high-precision abnormality perception operation and maintenance method based on drone inspection in the above embodiment, the present invention also provides a high-precision abnormality perception operation and maintenance system based on drone inspection, including:

[0086] The target indicator extraction module 11 is used to set a plurality of inspection targets according to the inspection target area, and perform feature mining on the plurality of inspection targets in turn to determine a plurality of target inspection indicator sets.

[0087] The inspection plan execution module 12 is used for the target UAV to take off regularly according to the preset inspection plan through the integrated take-off, landing and charging platform, and to inspect and monitor the multiple inspection targets based on the multiple target inspection indicator sets, wherein the target UAV is equipped with a high-precision sensor group.

[0088] The inspection data acquisition module 13 is used to obtain multiple target inspection data sets by using the high-precision sensor group inspection, perform abnormal value identification and data preprocessing on the multiple target inspection data sets according to the multiple target inspection indicator sets, and generate multiple target detection data sets.

[0089] The anomaly perception module 14 is used to build a multi-source heterogeneous anomaly perception algorithm model library based on the multiple inspection target training, call the multi-source heterogeneous anomaly perception algorithm model library to perform matching analysis on the multiple target detection data sets respectively, and obtain multiple target anomaly perception reports.

[0090] The decision management module 15 is used to transmit the multiple target abnormality perception reports to the operation and maintenance center for strategy analysis and optimization, obtain regional linkage operation and maintenance strategy parameters, and perform abnormal operation and maintenance control on the inspection target area based on the regional linkage operation and maintenance strategy parameters.

[0091] Among them, the target indicator extraction module 11 includes:

[0092] The characteristic analysis and basic characteristic data set acquisition unit is used to perform characteristic analysis on the multiple inspection targets in sequence to acquire multiple target basic characteristic data sets, wherein the multiple target basic characteristic data sets include location, model specification, application function and historical maintenance record.

[0093] The risk impact assessment and risk level information determination unit is used to perform risk impact assessment on the multiple inspection targets based on the multiple target basic characteristic data sets to obtain multiple target risk level information.

[0094] The feature mining depth determination unit is used to determine multiple target feature mining depths according to the multiple target risk level information.

[0095] The associated indicator mining and inspection indicator set generation unit is used to perform associated indicator mining on the multiple target basic characteristic data sets based on the multiple target feature mining depths to obtain the multiple target inspection indicator sets.

[0096] In some implementations, the associated indicator mining and inspection indicator set generation unit in the target indicator extraction module 11 includes:

[0097] The associated indicator mining and cluster marking unit is used to perform associated indicator mining on the multiple target basic characteristic data sets respectively to obtain multiple target associated indicator sets, and cluster mark the multiple target associated indicator sets to obtain multiple target classification indicator cluster sets.

[0098] The classification indicator cluster correlation analysis unit is used to perform correlation analysis on the multiple target classification indicator cluster sets based on the multiple target basic characteristic data sets to obtain multiple indicator cluster correlation coefficient sets.

[0099] The primary indicator screening unit is used to perform primary indicator screening on the multiple target classification indicator cluster sets according to the multiple target feature mining depths and the multiple indicator cluster correlation coefficient sets to obtain multiple target feature indicator cluster sets.

[0100] The feature indicator correlation analysis and secondary indicator screening unit is used to perform correlation analysis and secondary indicator screening on each feature indicator in the multiple target feature indicator cluster sets based on the multiple target feature mining depths to obtain the multiple target inspection indicator sets.

[0101] In some embodiments, the inspection data collection module 13 includes:

[0102] The abnormality identification factor setting and target data abnormality identification unit is used to set multiple target data abnormality identification factors according to the multiple target inspection indicator sets, and the multiple target data abnormality identification factors include data consistency, noise data and range rationality.

[0103] The outlier recognition and abnormal data set acquisition unit is used to perform outlier recognition on the multiple target inspection data sets based on the multiple target data abnormality recognition factors to obtain multiple target abnormal data sets.

[0104] The abnormal data preprocessing program mapping and data cleaning unit is used to map and obtain multiple target abnormal data preprocessing programs according to the multiple target data abnormality identification factors, and perform data cleaning on the multiple target abnormal data sets based on the multiple target abnormal data preprocessing programs to obtain multiple available target data sets.

[0105] The sensor experience error correction and target detection data set generation unit is used to perform feedback correction on the multiple available target data sets based on the sensor experience error factor to generate the multiple target detection data sets.

[0106] In some embodiments, the abnormality sensing module 14 includes:

[0107] The multi-source heterogeneous data collection and abnormal data set acquisition unit is used to perform multi-source heterogeneous data collection based on the multiple target inspection indicator sets to obtain multiple target multi-source heterogeneous abnormal data sets.

[0108] The anomaly perception feature extraction unit is used to extract anomaly perception features from the multiple target multi-source heterogeneous anomaly data sets to obtain multiple target anomaly perception data feature sets.

[0109] The feature algorithm selection and anomaly model training unit is used to perform feature algorithm selection and anomaly model training on the multiple target multi-source heterogeneous indicator data sets according to the multiple target anomaly perception data feature sets to obtain multiple target anomaly recognition model sets.

[0110] The anomaly recognition model integration unit is used to integrate the multiple target anomaly recognition model sets based on the multiple inspection targets to build the multi-source heterogeneous anomaly perception algorithm model library.

[0111] In some implementations, the multi-source heterogeneous data collection and abnormal data set acquisition unit in the abnormality perception module 1 includes:

[0112] The abnormal database mining and correlation collection unit is used to mine and obtain the regional inspection abnormal database, and perform first-order correlation collection in the regional inspection abnormal database based on the multiple target inspection indicator sets to obtain multiple first-order correlation indicator data sets.

[0113] The second-order correlation acquisition unit is used to perform second-order correlation acquisition in the regional inspection anomaly database based on the multiple first-order correlation indicator data sets to obtain multiple second-order correlation indicator data sets.

[0114] A multi-order correlation collection unit is used to traverse the correlation and collect N-1 order indicator data, where 2≤N≤4, based on the multiple first-order correlation indicator data sets, the multiple second-order correlation indicator data sets, and up to the N-1 order indicator data, to combine and obtain the multiple target multi-source heterogeneous anomaly data sets.

[0115] In some embodiments, the decision management module 15 includes:

[0116] The abnormality perception report matching and operation and maintenance strategy acquisition unit is used to perform feature matching with the abnormal operation and maintenance strategy library based on the abnormality perception reports of the multiple targets, and obtain abnormal operation and maintenance strategies applicable to the multiple targets.

[0117] An operation and maintenance strategy construction and parameter parsing unit is used to apply abnormal operation and maintenance strategies according to the multiple targets, construct multiple abnormal operation and maintenance strategy spaces, and perform parameter parsing on the multiple target abnormal perception reports according to the multiple abnormal operation and maintenance strategy spaces to obtain multiple target operation and maintenance strategy parameter thresholds.

[0118] The parameter global optimization unit is used to perform global optimization of parameters within the multiple target operation and maintenance strategy parameter thresholds according to a preset search step size to obtain multiple target operation and maintenance strategy parameters.

[0119] The linkage integration and regional linkage operation and maintenance strategy generation unit is used to link and integrate the multiple target operation and maintenance strategy parameters according to the drone inspection sequence to obtain the regional linkage operation and maintenance strategy parameters.

[0120] It should be understood that the embodiments mentioned in this specification focus on their differences from other embodiments. The specific embodiments in the aforementioned embodiment one are also applicable to the high-precision abnormality perception operation and maintenance system based on drone inspection described in embodiment two. For the sake of brevity of the specification, they will not be further elaborated here.

[0121] It should be understood that the embodiments disclosed in the present invention and the above description can enable those skilled in the art to use the present invention to implement the present invention. At the same time, the present invention is not limited to the above-mentioned embodiments. It should be understood that those skilled in the art can still modify the technical solutions recorded in the above-mentioned embodiments, or replace some of the technical features therein by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention.

Claims

1. A high-precision abnormality perception operation and maintenance method based on drone inspection, characterized in that: The method comprises: According to the inspection target area, multiple inspection targets are set, and feature mining is performed on the multiple inspection targets in turn to determine multiple target inspection indicator sets; The target UAV takes off at a fixed time according to a preset inspection plan through the integrated take-off, landing and charging platform, and performs inspection and monitoring on the multiple inspection targets based on the multiple target inspection indicator sets, wherein the target UAV is equipped with a high-precision sensor group; Acquire multiple target inspection data sets by using the high-precision sensor group inspection, perform outlier identification and data preprocessing on the multiple target inspection data sets according to the multiple target inspection indicator sets, and generate multiple target detection data sets; Building a multi-source heterogeneous anomaly perception algorithm model library based on the multiple inspection target training, calling the multi-source heterogeneous anomaly perception algorithm model library to perform matching analysis on the multiple target detection data sets respectively, and obtaining multiple target anomaly perception reports; The multiple target abnormality perception reports are transmitted to the operation and maintenance center for strategy analysis and optimization, to obtain regional linkage operation and maintenance strategy parameters, and based on the regional linkage operation and maintenance strategy parameters, abnormal operation and maintenance control is performed on the inspection target area.

2. The high-precision abnormality perception operation and maintenance method based on drone inspection as claimed in claim 1 is characterized in that: Determining multiple target inspection indicator sets includes: Performing characteristic analysis on the multiple inspection targets in turn to obtain multiple target basic characteristic data sets, wherein the multiple target basic characteristic data sets include location, model specification, application function and historical maintenance record; Perform risk impact assessment on the multiple inspection targets based on the multiple target basic characteristic data sets to obtain multiple target risk level information; Determining multiple target feature mining depths according to the multiple target risk level information; Based on the multiple target feature mining depths, the multiple target basic characteristic data sets are respectively mined for associated indicators to obtain the multiple target inspection indicator sets.

3. The high-precision abnormality perception operation and maintenance method based on drone inspection as claimed in claim 2 is characterized in that: The obtaining of the plurality of target inspection indicator sets comprises: Performing associated indicator mining on the multiple target basic characteristic data sets respectively to obtain multiple target associated indicator sets, and performing cluster marking on the multiple target associated indicator sets to obtain multiple target classification indicator cluster sets; Based on the multiple target basic characteristic data sets, correlation analysis is performed on the multiple target classification indicator cluster sets to obtain multiple indicator cluster correlation coefficient sets; According to the multiple target feature mining depths and the multiple indicator cluster correlation coefficient sets, the multiple target classification indicator cluster sets are screened for primary indicators to obtain multiple target feature indicator cluster sets; Based on the multiple target feature mining depths, correlation analysis and secondary indicator screening are performed on each feature indicator in the multiple target feature indicator cluster sets to obtain the multiple target inspection indicator sets.

4. The high-precision abnormality perception operation and maintenance method based on drone inspection as claimed in claim 1 is characterized in that: The generating of multiple target detection data sets includes: According to the multiple target inspection indicator sets, multiple target data anomaly identification factors are set, wherein the multiple target data anomaly identification factors include data consistency, noise data, and range rationality; Based on the multiple target data anomaly identification factors, outlier identification is performed on the multiple target inspection data sets to obtain multiple target anomaly data sets; According to the multiple target data anomaly identification factors, mapping obtains multiple target anomaly data preprocessing programs, and performing data cleaning on the multiple target anomaly data sets based on the multiple target anomaly data preprocessing programs to obtain multiple available target data sets; Feedback correction is performed on the multiple available target data sets based on sensor empirical error factors to generate the multiple target detection data sets.

5. The high-precision abnormality perception operation and maintenance method based on drone inspection as claimed in claim 1 is characterized in that: The construction of a multi-source heterogeneous anomaly perception algorithm model library includes: Perform multi-source heterogeneous data collection based on the multiple target inspection indicator sets to obtain multiple target multi-source heterogeneous abnormal data sets; Extracting anomaly perception features from the multiple target multi-source heterogeneous anomaly data sets to obtain multiple target anomaly perception data feature sets; According to the multiple target abnormal perception data feature sets, feature algorithm selection and abnormal model training are performed on the multiple target multi-source heterogeneous indicator data sets to obtain multiple target abnormality recognition model sets; The multiple target anomaly recognition model sets are integrated based on the multiple inspection targets to construct the multi-source heterogeneous anomaly perception algorithm model library.

6. The high-precision abnormality perception operation and maintenance method based on drone inspection as claimed in claim 5 is characterized in that: The obtaining of multiple target multi-source heterogeneous abnormal data sets includes: Mining and acquiring a regional inspection anomaly database, performing first-order correlation collection in the regional inspection anomaly database based on the multiple target inspection indicator sets, and obtaining multiple first-order correlation indicator data sets; Based on the multiple first-order correlation index data sets, second-order correlation collection is performed in the regional inspection anomaly database to obtain multiple second-order correlation index data sets; Until N-1 order indicator data is traversed and associated, where 2≤N≤4, the multiple first-order associated indicator data sets, the multiple second-order associated indicator data sets, until the N-1 order indicator data, are combined to obtain the multiple target multi-source heterogeneous anomaly data sets.

7. The high-precision abnormality perception operation and maintenance method based on drone inspection as claimed in claim 1 is characterized in that: The obtaining of regional linkage operation and maintenance strategy parameters includes: Based on the abnormal perception reports of the multiple targets, feature matching is performed with the abnormal operation and maintenance strategy library to obtain abnormal operation and maintenance strategies applicable to the multiple targets; According to the multiple targets, the abnormal operation and maintenance strategies are applied, multiple abnormal operation and maintenance strategy spaces are constructed, and the parameters of the multiple target abnormal perception reports are parsed according to the multiple abnormal operation and maintenance strategy spaces to obtain multiple target operation and maintenance strategy parameter thresholds; Performing global optimization of parameters within the multiple target operation and maintenance strategy parameter thresholds according to a preset search step length to obtain multiple target operation and maintenance strategy parameters; The multiple target operation and maintenance strategy parameters are linked and integrated according to the drone inspection sequence to obtain the regional linkage operation and maintenance strategy parameters.

8. High-precision abnormality perception operation and maintenance system based on drone inspection, characterized by: The system is used to execute the high-precision abnormality perception operation and maintenance method based on drone inspection according to any one of claims 1 to 7, and the system includes: A target indicator extraction module, which is used to set multiple inspection targets according to the inspection target area, and perform feature mining on the multiple inspection targets in turn to determine multiple target inspection indicator sets; An inspection plan execution module, wherein the target UAV takes off at a scheduled time according to a preset inspection plan through the integrated take-off, landing and charging platform, and performs inspection and monitoring on the multiple inspection targets based on the multiple target inspection indicator sets, wherein the target UAV is equipped with a high-precision sensor group; An inspection data acquisition module, the inspection data acquisition module is used to obtain multiple target inspection data sets by using the high-precision sensor group inspection, perform outlier recognition and data preprocessing on the multiple target inspection data sets according to the multiple target inspection indicator sets, and generate multiple target detection data sets; An anomaly perception module, the anomaly perception module is used to build a multi-source heterogeneous anomaly perception algorithm model library based on the multiple inspection target training, call the multi-source heterogeneous anomaly perception algorithm model library to perform matching analysis on the multiple target detection data sets respectively, and obtain multiple target anomaly perception reports; A decision management module is used to transmit the multiple target abnormality perception reports to the operation and maintenance center for strategy analysis and optimization, obtain regional linkage operation and maintenance strategy parameters, and perform abnormal operation and maintenance control on the inspection target area based on the regional linkage operation and maintenance strategy parameters.

Images