Pipeline fault intelligent identification and positioning system based on unmanned aerial vehicle technology

Abstract

The application relates to the technical field of pipeline detection, and discloses a pipeline fault intelligent identification and positioning system based on unmanned aerial vehicle technology, which comprises a data acquisition end, a data processing end and a decision output end; the unmanned aerial vehicle platform loaded with various sensors automatically patrols and collects along a preset path, and comprehensively utilizes image, point cloud, sound wave, gas and temperature field data to construct a multi-dimensional and multi-parameter cooperative sensing system, so that the limitations of traditional manual inspection and single-sensor unmanned aerial vehicle monitoring in efficiency, coverage range and sensing dimension are overcome. A multi-dimensional mapping relationship is established for multi-source sensing data, and key feature parameters with unified space-time labels are extracted therefrom, so that comprehensive detection of various potential faults such as pipeline surface corrosion, slight leakage and structure deformation is realized, early detection capability of hidden faults is improved, and safety hazards are reduced.

Description

Technical Field

[0001] This invention relates to the field of pipeline inspection technology, specifically to an intelligent pipeline fault identification and location system based on unmanned aerial vehicle (UAV) technology. Background Technology

[0002] As vital carriers for transmitting information, transporting energy, and handling sewage and flood control, pipelines are an indispensable part of the infrastructure in daily life. With the rapid development of modern cities, underground pipelines are playing an increasingly important role. However, incidents involving underground pipelines, such as leaks, explosions, and road collapses, have also occurred. The high-rise construction of various facilities has made pipeline structures more complex, and the fact that most pipelines are buried underground has brought certain difficulties to pipeline fault detection.

[0003] Currently, in the practice of using drones for pipeline inspection, traditional pipeline inspection mainly relies on manual inspection or data collection by drones equipped with a single sensor. This method is not only inefficient and has a limited coverage, but it is also difficult to perceive potential faults such as pipeline surface corrosion, minor leaks, and deformation around the clock and from multiple angles. As a result, hidden faults cannot be detected in time, posing serious safety hazards.

[0004] Therefore, a pipeline fault intelligent identification and location system based on UAV technology is proposed to solve the above problems. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides an intelligent pipeline fault identification and location system based on unmanned aerial vehicle (UAV) technology, which solves the problem mentioned in the background that hidden faults cannot be detected in a timely manner, posing serious safety hazards.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a pipeline fault intelligent identification and location system based on UAV technology, the system comprising a data acquisition terminal, a data processing terminal, and a decision output terminal; The data acquisition terminal is used to collect multi-source sensing data and its observation sequence information of the pipeline by a drone equipped with multiple sensors according to a preset path and interval, and send it to the data processing terminal. The multi-source sensing data includes image, point cloud, sound wave, gas concentration and temperature field data, and the observation sequence information includes flight log sequence number, collection timestamp and spatial coordinate sequence. The data processing terminal is used to receive the multi-source sensing data and observation sequence information, construct a multi-dimensional mapping relationship between the observation sequence information and the acquisition time, spatial location and environmental parameters, and based on this mapping relationship, extract key feature parameters with spatiotemporal labels from the multi-source sensing data, and perform cross-scale correlation analysis on the feature parameters to jointly determine the fault type and reversely deduce the fault state time. The decision output terminal is used to receive the fault type and fault status time. By comparing the time interval between the fault status time and the current time with a preset threshold, the fault level is determined as a hot fault, a warm fault, or a cold fault. Different fault pipeline areas of different levels are differentiated and spatially located in the geographic information system.

[0007] Preferably, the data acquisition terminal includes a flight control module, a sensor array module, and a data encapsulation module; The flight control module includes a path planning unit, a state synchronization unit, and a sequence generation unit; The path planning unit is used to plan the preset cruise path of the UAV based on the geographical information of the pipeline to be monitored, and to set the time interval for data collection. The state synchronization unit is used to synchronize and record environmental parameters, UAV platform flight status parameters and sensor working status parameters in real time at each acquisition moment during UAV flight. The environmental parameters include wind speed, light intensity and ambient temperature, and the platform flight status parameters include flight altitude, flight speed and attitude angle. The sequence generation unit is used to generate a unique observation sequence identifier for each data acquisition event. The observation sequence identifier includes the track point number, image frame sequence number, laser point cloud timestamp sequence number, and acoustic sampling sequence number.

[0008] Preferably, the sensor array module includes an optical sensing unit, a point cloud sensing unit, an acoustic wave sensing unit, a gas sensing unit, and a temperature sensing unit. The optical sensing unit is used to acquire optical and infrared image data of the pipeline and its surrounding environment through a high-resolution optical camera and infrared thermal imager mounted on the UAV. The point cloud sensing unit is used to collect three-dimensional point cloud data of pipelines and their ancillary facilities through the lidar equipment carried by the UAV. The acoustic wave sensing unit is used to collect acoustic wave signals and vibration signal data generated during pipeline operation through acoustic wave sensors and acoustic emission sensors carried by the UAV. The gas sensing unit is used to collect concentration data of specific gas components in the air around the pipeline using a gas sensor carried by the UAV. The temperature sensing unit is used to collect temperature field distribution data on the pipeline surface through infrared temperature measurement equipment and temperature sensor array carried by the UAV.

[0009] Preferably, the data encapsulation module includes a data receiving unit, a timestamp binding unit, and a data packet generation unit; The data receiving unit is used to receive multi-source sensing data from the sensor array module in real time, as well as observation sequence identifiers, environmental parameters, and UAV platform status parameters from the flight control module. The timestamp binding unit is used to bind the corresponding timestamp during the acquisition process to each set of multi-source sensing data and its corresponding observation sequence identifier received; The data packet generation unit is used to integrate and encapsulate multi-source sensing data with timestamps, observation sequence identifiers, corresponding UAV spatial coordinates, environmental parameters, and platform status parameters into a sensing data packet with a unified sequence identifier, and send it to the data processing terminal.

[0010] Preferably, the data processing terminal includes a mapping construction module, a feature parsing module, and an intelligent association module; The mapping construction module includes a sequence parsing unit, a spatiotemporal association unit, and a context binding unit; The sequence parsing unit is used to parse the received sensing data packets and extract the observation sequence identifier, collection timestamp and UAV spatial coordinates. The spatiotemporal correlation unit is used to construct a forward mapping chain from the sequence identifier to the acquisition timestamp and then to the spatial coordinates, using the observation sequence identifier as an index, and to establish a corresponding reverse query index. The context binding unit is used to further bind the acquisition timestamp with environmental parameters, platform flight status parameters, and sensor operating status parameters to form a multi-dimensional mapping relationship network for subsequent comprehensive analysis.

[0011] Preferably, the feature parsing module includes an image analysis unit, a point cloud analysis unit, a voiceprint analysis unit, a gas analysis unit, and a thermal imaging analysis unit; The image analysis unit is used to perform color analysis, texture analysis and contour recognition on optical and infrared images based on the multi-dimensional mapping relationship, extract color abnormality areas, texture abnormality areas, structural cracks, leakage reflection and surface corrosion feature parameters in the images, and assign a corresponding acquisition timestamp and spatial location label to each feature parameter. The point cloud analysis unit is used to perform registration and comparison analysis on the three-dimensional point cloud data based on the multi-dimensional mapping relationship, calculate the deformation, displacement, settlement and geometric anomaly index of the pipeline, and assign a corresponding collection timestamp and spatial location label to each deformation index. The acoustic signature analysis unit is used to perform time-frequency domain transformation and feature extraction on the acoustic signal based on the multi-dimensional mapping relationship, identify abnormal frequency components, energy mutations and acoustic emission features of specific modes in the signal, and assign a corresponding acquisition timestamp and spatial location label to each acoustic feature. The gas analysis unit is used to perform component analysis and threshold comparison on the concentration data collected by the gas sensor based on the multi-dimensional mapping relationship, identify the concentration exceedance and leakage characteristics of specific gas components, and assign a corresponding collection timestamp and spatial location label to each gas anomaly characteristic. The thermal imaging analysis unit is used to perform temperature field analysis and region segmentation on the thermal imaging data acquired by the infrared thermal imager based on the multi-dimensional mapping relationship, extract temperature anomaly regions, temperature gradient anomalies and hot spot features, and assign a corresponding acquisition timestamp and spatial location label to each thermal feature.

[0012] Preferably, the intelligent association module includes a data aggregation unit, a pattern matching unit, and a time inversion unit; The data aggregation unit is used to aggregate and align key feature parameters with spatiotemporal labels extracted from different sensors at different times in the same pipeline area, with spatial location as the core, to form a feature parameter set for that location. The pattern matching unit is used to match and analyze the set of feature parameters with a preset fault mode knowledge base. The fault mode knowledge base stores multi-source feature combination templates corresponding to various fault types. Through matching, the fault type corresponding to the pipeline area is identified. The fault types include minor leaks, gas leaks, material aging, external corrosion, structural cracks, and mechanical damage. The time inversion unit is used to perform reverse derivation of the fault occurrence time based on the multi-dimensional mapping relationship and the feature parameter set. Specifically, it traverses the anomaly scores of each feature parameter in the feature parameter set in sequence on the time axis, identifies the time node when the anomaly score first exceeds the preset threshold, and determines the time node as the fault state time of the pipeline area.

[0013] Preferably, the decision output terminal includes a level determination module, an early warning response module, and a visualization presentation module; The grade determination module includes an interval calculation unit, a threshold comparison unit, and a dynamic weighting unit; The interval calculation unit is used to calculate the time interval between the fault status time output by the data processing terminal and the current system time; The threshold comparison unit is used to compare the time interval with a preset first threshold and a second threshold, and to classify the fault level according to the comparison result; The dynamic weighting unit is used to dynamically correct and adjust the initially determined fault level based on threshold comparison, combined with historical fault frequency data and environmental stress condition data of the faulty pipeline area.

[0014] Preferably, the early warning response module includes a thermal fault handling unit, a temperature fault management unit, and a cold fault archiving unit; The thermal fault handling unit is used to automatically trigger a high-level emergency response mechanism for pipeline areas that are determined to be thermal faults. The mechanism includes sending real-time audible and visual alarms to the operation and maintenance terminal, highlighting and flashing marks on the GIS map, automatically generating emergency inspection task work orders, and planning the optimal path to the fault point. The temperature fault management unit is used to include pipeline areas identified as having temperature faults in the system's priority inspection list, increase their task priority in subsequent periodic inspection plans, and generate potential risk investigation and suggestion reports. The cold fault archiving unit is used to automatically archive and store pipeline areas identified as cold faults, and generate historical fault analysis reports containing fault location, type, status time, and evolution process analysis, providing data support for optimizing long-term pipeline maintenance strategies.

[0015] Preferably, the visualization module includes a GIS platform integration unit, a differentiated rendering unit, and an interactive analysis unit; The GIS platform integration unit is used to import the fault information set, which includes the fault spatial location, fault type, fault level and fault status time, into the geographic information system platform, and perform spatial registration and overlay with the underlying pipeline network base map. The differentiated rendering unit is used to render and display different visual identification strategies on the GIS map according to the fault level: a first preset visual style is used to identify thermal faults, a second preset visual style is used to identify warm faults, and a third preset visual style is used to identify cold faults. The difference in the visual style is reflected in the shape, color, brightness and dynamic effect of the identification symbols. The interactive analysis unit provides a map interaction interface, enabling maintenance personnel to click and query fault icons displayed on the GIS map, filter regions, and perform multi-dimensional statistics. In response to user operations, it dynamically displays detailed attribute information of the selected fault, associated multi-source feature evidence, and system-recommended handling measures.

[0016] Compared with existing technologies, the present invention provides a pipeline fault intelligent identification and location system based on UAV technology, which has the following beneficial effects: 1. In this invention, a multi-sensor drone platform automatically cruises along a preset path to collect data. By comprehensively utilizing image, point cloud, sound wave, gas, and temperature field data, a multi-dimensional, multi-parameter collaborative sensing system is constructed, overcoming the limitations of traditional manual inspections and single-sensor drone monitoring in terms of efficiency, coverage, and sensing dimensions. By establishing multi-dimensional mapping relationships between multi-source sensing data and extracting key feature parameters with unified spatiotemporal labels, comprehensive detection of various potential faults such as pipeline surface corrosion, minor leaks, and structural deformation is achieved. This enhances the early detection capability of hidden faults and reduces safety hazards.

[0017] 2. In this invention, by performing cross-scale correlation analysis and joint judgment on multi-source feature parameters with spatiotemporal labels, weak fault signs of different physical quantities can be correlated and verified in spatiotemporal terms, thereby achieving diagnosis and comprehensive judgment of complex fault types. This decision-making mechanism based on multi-source information fusion improves the accuracy and reliability of fault type identification and location, and reduces the probability of false alarms and missed alarms.

[0018] 3. In this invention, by comparing the fault state time with the current time to classify faults into hot, warm, and cold levels, and combining historical data with environmental factors for dynamic weighting, intelligent and quantitative assessment of the urgency of faults is achieved. Based on this, the system can automatically trigger differentiated high-level emergency response mechanisms according to different fault levels, and display the corresponding differentiated visual representation on a GIS map. This mechanism shifts the operation and maintenance response from a uniform mode to differentiated classification, ensuring rapid emergency response to sudden high-risk faults, while optimizing the allocation of operation and maintenance resources to medium- and low-risk events, thereby improving the intelligence level and proactive safety assurance capabilities of pipeline operation and maintenance management. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the intelligent pipeline fault identification and location system based on UAV technology according to the present invention. Figure 2 This is a functional module structure diagram of a pipeline fault intelligent identification and location system based on UAV technology according to the present invention. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] For specific implementation examples, please refer to: Figure 1-2A pipeline fault intelligent identification and location system based on UAV technology, the system includes a data acquisition end, a data processing end and a decision output end; The data acquisition end is used to collect multi-source sensing data and its observation sequence information of the pipeline by using a drone equipped with multiple sensors according to a preset path and interval, and send it to the data processing end. The multi-source sensing data includes images, point clouds, sound waves, gas concentration and temperature field data, and the observation sequence information includes flight log sequence number, collection timestamp and spatial coordinate sequence. The data processing unit receives multi-source sensing data and observation sequence information, constructs a multi-dimensional mapping relationship between the observation sequence information and the acquisition time, spatial location, and environmental parameters. Based on this mapping relationship, it extracts key feature parameters with spatiotemporal labels from the multi-source sensing data and performs cross-scale correlation analysis on the feature parameters to jointly determine the fault type and inversely deduce the fault state time. The cross-scale correlation analysis executes the following steps in sequence: Spatial scale association: The visual features extracted by the image analysis unit, the geometric deformation features extracted by the point cloud analysis unit, and the acoustic features extracted by the voiceprint analysis unit at the same geographical location are aligned and superimposed to form a multi-source feature vector for that location; Time-scale correlation: Based on multi-dimensional mapping relationships, trace and compare the multi-source feature vectors corresponding to the same pipeline location in different historical inspection cycles, calculate the changing trend and rate of each feature parameter over time, and for any feature parameter... It is in two adjacent inspection cycles and Rate of change between Specifically, it is calculated using the following formula: ; in, For feature parameters The rate of change, Indicates time Collected feature parameters The value; Logical correlation analysis: When the characteristics of "excessive gas concentration" and "abnormal infrared temperature" appear simultaneously at the same location, the joint analysis is "leakage-related fault". When the characteristics of "geometric deformation" and "acoustic emission at a specific frequency" appear simultaneously, the joint analysis is "structural damage-related fault". The decision output end is used to receive the fault type and fault status time. By comparing the time interval between the fault status time and the current time with a preset threshold, the fault level is determined as a hot fault, a warm fault, or a cold fault. Different fault pipeline areas of different levels are differentiated and spatially located in the geographic information system.

[0022] The data acquisition unit includes a flight control module, a sensor array module, and a data encapsulation module; The flight control module includes a path planning unit, a state synchronization unit, and a sequence generation unit; The path planning unit is used to plan the preset cruise path of the UAV based on the geographical information of the pipeline to be monitored, and to set the time interval for data collection. The state synchronization unit is used to synchronize and record environmental parameters, UAV platform flight status parameters and sensor working status parameters in real time at each acquisition moment during UAV flight. Environmental parameters include wind speed, light intensity and ambient temperature, and platform flight status parameters include flight altitude, flight speed and attitude angle. The sequence generation unit is used to generate a unique observation sequence identifier for each data acquisition event. The observation sequence identifier includes the track point number, image frame sequence number, laser point cloud timestamp sequence number, and acoustic sampling sequence number.

[0023] The sensor array module includes an optical sensing unit, a point cloud sensing unit, an acoustic wave sensing unit, a gas sensing unit, and a temperature sensing unit. The optical sensing unit is used to acquire optical and infrared image data of pipelines and their surrounding environment through a high-resolution optical camera and infrared thermal imager carried by the drone. The point cloud sensing unit is used to collect three-dimensional point cloud data of pipelines and their ancillary facilities through the lidar equipment carried by the UAV; The acoustic sensing unit is used to collect acoustic and vibration signal data generated during pipeline operation through acoustic sensors and acoustic emission sensors carried by the UAV. The gas sensing unit is used to collect concentration data of specific gas components in the air around the pipeline using gas sensors carried by the drone; The temperature sensing unit is used to collect temperature field distribution data on the pipeline surface through infrared temperature measurement equipment and temperature sensor array carried by the drone.

[0024] The data encapsulation module includes a data receiving unit, a timestamp binding unit, and a data packet generation unit; The data receiving unit is used to receive multi-source sensing data from the sensor array module in real time, as well as observation sequence identifiers, environmental parameters, and UAV platform status parameters from the flight control module; The timestamp binding unit is used to bind the corresponding timestamp during the acquisition process to each set of multi-source sensing data and its corresponding observation sequence identifier. The data packet generation unit is used to integrate and encapsulate multi-source sensing data with timestamps, observation sequence identifiers, corresponding UAV spatial coordinates, environmental parameters, and platform status parameters into a sensing data packet with a unified sequence identifier, and send it to the data processing end.

[0025] The data processing module includes a mapping construction module, a feature parsing module, and an intelligent association module; The mapping building block includes a sequence parsing unit, a spatiotemporal association unit, and a context binding unit; The sequence parsing unit is used to parse the received sensing data packets and extract the observation sequence identifier, acquisition timestamp, and UAV spatial coordinates. The spatiotemporal correlation unit is used to construct a forward mapping chain from the observation sequence identifier to the acquisition timestamp and then to the spatial coordinates, using the observation sequence identifier as an index, and to establish a corresponding reverse query index, specifically including: Create a forward mapping table: Create a mapping table in the storage system with the observation sequence identifier as the primary key. Each record in the table contains a collection timestamp field that uniquely corresponds to the identifier, and a representative spatial location field calculated based on the UAV spatial coordinate sequence corresponding to that timestamp. Construct a forward mapping chain: For each observation sequence identifier, based on the records in the mapping table, form a unidirectional data chain with a pointing relationship from the observation sequence identifier to its unique acquisition timestamp, and then to its determined spatial location. Establish a time-reverse index: Using the collection timestamp as the query key, establish a reverse index structure that can quickly retrieve and return all observation sequence identifiers and their associated spatial locations corresponding to the data collected at that time point; Establish a spatial reverse index: Using geospatial coordinates or coordinate ranges as query conditions, establish a reverse index structure that can quickly retrieve and return all observation sequence identifiers and their associated collection timestamps falling within the spatial range; The context binding unit is used to further bind the acquisition timestamp with environmental parameters, platform flight status parameters, and sensor operating status parameters, forming a multi-dimensional mapping network for subsequent comprehensive analysis, specifically including: A basic mapping table is created using the observation sequence identifier as the primary key, with fields including collection timestamp, longitude, latitude, and altitude. Extended context fields: In the base mapping table, add corresponding environmental parameter fields, UAV platform flight status parameter fields, and sensor working status parameter fields for each record; Establish an inverse index: Using the geospatial coordinate range and the collection timestamp range as joint conditions, establish an inverse index database that can quickly query and retrieve all corresponding observation sequence identifiers and all associated fields.

[0026] The feature parsing module includes an image analysis unit, a point cloud analysis unit, a voiceprint analysis unit, a gas analysis unit, and a thermal imaging analysis unit; The image analysis unit performs color analysis, texture analysis, and contour recognition on optical and infrared images based on multi-dimensional mapping relationships. It extracts feature parameters of color anomalies, texture anomalies, structural cracks, leakage reflections, and surface corrosion from the images, and assigns a corresponding acquisition timestamp and spatial location label to each feature parameter. Specifically, this includes: For optical images, they are converted to the HSV color space. A statistical histogram of hue and saturation within a pre-defined detection area on the pipeline surface is calculated and compared with the histogram of a standard sample using the Bartlett distance method. The calculation method is defined as follows: ; in, For Bach distance, and These represent the current region and the standard sample at the [number]th [year]. The probability value of each histogram interval. The total number of intervals, For distance measurement; when If the value exceeds the first preset value, it will be marked as a color abnormality area; For the same region, the local binary mode algorithm is used to extract texture features and calculate its texture feature vector. Texture feature vectors compared to standard samples The Euclidean distance is calculated as follows: ; in, For Euclidean distance, This is the texture feature vector of the current region. The texture feature vector of the standard sample. The dimension of the texture feature vector. The vector represents the first One portion, For distance measurement; when If the value exceeds the second set value, it will be marked as a texture abnormality area; The point cloud analysis unit is used to register and compare 3D point cloud data based on multi-dimensional mapping relationships, calculate pipeline deformation, displacement, settlement, and geometric anomaly indicators, and assign a corresponding acquisition timestamp and spatial location label to each deformation indicator. Specifically, this includes: Point cloud registration: The pipeline point cloud data acquired during the current inspection is registered with the pre-stored reference 3D point cloud model of the pipeline using an iterative nearest-point algorithm. The objective function of this algorithm is to minimize the registration error between corresponding points in the current point cloud and the reference point cloud, defined as: ; in, For registration error, For the current point cloud One point, It is the reference point in the cloud and Corresponding points It is a rotation matrix. It is a translation vector. It represents the number of corresponding point pairs; Distance Comparison: Calculate the nearest distance from each point in the current point cloud to the corresponding surface of the reference point cloud model, and generate a distance field. ; Deformation extraction: for range field Threshold segmentation is performed, separating values ​​whose distance exceeds a set tolerance threshold. The point cloud region is extracted, and the centroid displacement and average distance deviation of the point cloud in that region are calculated as deformation and displacement indicators. The acoustic signature analysis unit is used to perform time-frequency domain transformation and feature extraction on acoustic signals based on multi-dimensional mapping relationships, identify abnormal frequency components, energy abrupt changes and acoustic emission characteristics of specific modes in the signal, and assign a corresponding acquisition timestamp and spatial location label to each acoustic feature. The gas analysis unit is used to perform component analysis and threshold comparison on the concentration data collected by gas sensors based on multi-dimensional mapping relationships, identify the concentration exceedance and leakage characteristics of specific gas components, and assign a corresponding collection timestamp and spatial location label to each gas anomaly characteristic, specifically including: Concentration reading: Extracting the real-time concentration value of a specific gas component from the data of the gas sensing unit. ; Background concentration subtraction: Read the historical average background concentration of this monitoring point under fault-free conditions. ; Safety threshold setting: safe concentration threshold The settings are based on the background concentration and its statistical fluctuation range, and the calculation method is defined as follows: ; in, For safe concentration thresholds, This represents the historical average concentration. The sample standard deviation of historical background concentration data. The safety factor, which is preset based on the hazard level of the gas composition, is a constant greater than zero. Confirmation time setting: Preset confirmation time for concentration exceeding the standard The settings are based on the gas diffusion rate and system response requirements, and the calculation method is defined as follows: ; in, The preset time for confirming concentration exceeding the standard. The effective response spatial scale of the sensor, This represents the typical diffusion rate of the gas in air. The preset system de-jitter and confirmation buffer time; Anomaly detection: Calculate concentration deviation Its calculation method is defined as follows: ; in, The deviation between real-time concentration and background concentration. This represents the real-time concentration value of the gas. This represents the historical average concentration. when The concentration of this gas component continues to exceed the safe concentration threshold. Reaching the preset duration Then, characteristic parameters of gas concentration exceeding the standard are generated, and the start time of exceeding the standard is recorded; The thermal imaging analysis unit is used to perform temperature field analysis and region segmentation on the thermal imaging data acquired by the infrared thermal imager based on multi-dimensional mapping relationships. It extracts features such as temperature anomalies, temperature gradient anomalies, and hot spot characteristics, and assigns a corresponding acquisition timestamp and spatial location label to each thermal feature. Specifically, this includes: Temperature field calibration: Based on the ambient temperature parameters in the infrared image and multi-dimensional mapping relationship, the infrared temperature readings on the pipeline surface are calibrated by environmental reflection compensation. Zone Temperature Extraction: Delineate areas of interest at key locations on the pipeline surface and calculate the average temperature of those areas. and highest temperature ; Temperature difference analysis: Calculate the average temperature of the area of ​​interest. Average temperature of adjacent normal pipe wall region The difference Its calculation method is defined as follows: ; in, To monitor the temperature difference between the monitored area and the normal area, Pay attention to the average temperature of the area. This represents the average temperature of the adjacent normal pipe wall region. when Exceeding the preset temperature difference abnormality threshold Then, abnormal temperature field characteristic parameters are generated.

[0027] The intelligent association module includes a data aggregation unit, a pattern matching unit, and a time reversal unit; The data aggregation unit is used to aggregate and align key feature parameters with spatiotemporal labels extracted from different sensors at different times in the same pipeline area, with spatial location as the core, to form a set of feature parameters for that location. The pattern matching unit is used to perform matching analysis between the feature parameter set and a preset fault mode knowledge base. The fault mode knowledge base stores multi-source feature combination templates corresponding to various fault types. Through matching, the fault type corresponding to the pipeline area is identified. Fault types include minor leaks, gas leaks, material aging, external corrosion, structural cracks, and mechanical damage, specifically including: Feature vectorization: Normalize all feature parameters in the feature parameter set and concatenate them into a comprehensive feature vector in a predetermined order. ; Template library comparison: combining feature vectors The feature vector of each fault template in the fault mode knowledge base Similarity calculation is performed using the cosine similarity method, which is defined as follows: ; in, For cosine similarity, For the comprehensive feature vector, For the first in the Fault Mode and Effects Knowledge Base Feature vectors of each fault template Represents the magnitude of a vector; similarity The value range is [-1, 1]. The closer the value is to 1, the higher the similarity. The similarity calculation result will be compared with a matching confidence threshold. Compare the results to make a final judgment; Confidence threshold setting: Match confidence threshold The calculation method is defined based on the historical similarity calculation results of all known fault template vectors in the fault mode knowledge base with itself and other template vectors. ; in, To match the confidence threshold, and These are the sample mean similarity values ​​of all known correct matches in the knowledge base, and the standard deviation of the similarity distribution between correct and incorrect matches, respectively. The preset confidence level coefficient is a constant greater than zero. Fault determination: Select the fault template with the highest similarity calculation result. When their similarity value Exceeding the matching confidence threshold If so, the fault type corresponding to the current pipeline area is determined to be the fault type defined by the template; The time inversion unit is used to perform reverse derivation of the fault occurrence time based on multi-dimensional mapping relationships and a set of feature parameters. Its specific operation is as follows: on the time axis, it sequentially traverses the anomaly scores of each feature parameter in the feature parameter set, identifies the time node when the anomaly score first exceeds a preset threshold, and determines this time node as the fault state time of the pipeline area. Specifically, this includes: Score generation: For each feature parameter in the feature parameter set... According to its current value deviating from the preset normal range The degree of abnormality is calculated using a linear mapping method, resulting in an anomaly score between 0 and 100. Its calculation method is defined as follows: ; in, For anomaly score, This represents the current value of the feature parameter. This is the preset normal range for this feature parameter. This is the lower limit threshold of the normal range for this feature parameter. This is the upper limit threshold of the normal range for this feature parameter. and These are the upper and lower limits of the pre-defined feature parameter, used to linearly map values ​​outside the normal range to the (0,100] interval; Threshold setting: for each feature parameter Its static threshold This parameter is set based on anomaly scores from historical normal operating conditions, and its calculation method is defined as follows: ; in, For the first The static threshold of each feature parameter, and These are the characteristic parameters. Anomaly score corresponding to the historical normal operation cycle The sample mean and sample standard deviation, The preset confidence coefficient is a constant greater than zero. Time series construction: Based on the multi-dimensional mapping relationship, the anomaly scores of all feature parameters are arranged in chronological order according to their corresponding collection timestamps to form a multi-dimensional anomaly score time series for the pipeline location; First threshold determination: In the time series, the first time node that meets the following conditions is identified. Time of being identified as a fault state: A characteristic parameter exists, and its anomaly score is given. And its score at all times prior to that point in time. .

[0028] The decision output includes a level determination module, an early warning response module, and a visualization presentation module; The rating determination module includes an interval calculation unit, a threshold comparison unit, and a dynamic weighting unit; The interval calculation unit is used to calculate the time interval between the fault status time output by the data processing terminal and the current system time; The threshold comparison unit is used to compare the time interval with preset first and second thresholds, and classify the fault level based on the comparison result. The fault level determination logic is as follows: Calculate fault state time With the current system time The time interval between : ; in, This is the time interval between the fault state time and the current time. The current system time. The fault state time; when If so, it is determined to be a thermal fault. If so, it is determined to be a temperature fault. If the condition is as described above, it is determined to be a cold fault; among which, and The preset time threshold, and ; The dynamic weighting unit is used to dynamically correct and adjust the initially determined fault level based on threshold comparison, combined with historical fault frequency data and environmental stress condition data of the faulty pipeline area. Specifically, it includes: Historical weighting factor calculation: Query the number of times the same type of failure has occurred in the pipeline area in the past year. Calculate historical weighting factors Its calculation method is defined as follows: ; in, This is the historical weighting factor, with a maximum value of 1.0; Environmental stress factor calculation: Obtain environmental stress condition data at the time of the fault. When the wind speed exceeds the warning value, set the environmental stress factor. If the value is 1.2, then it is 1.0; Level-based weighting: Prioritizing the basic response based on the initially determined fault level. Multiplied by historical weighting factor and environmental stress factor This yields the final dynamic response priority. Its calculation method is defined as follows: ; in, For dynamic response priority, Based on response priority, Historical weighting factor Environmental stress factor; in accordance with The numerical range is used to adjust the fault level for upgrading, maintaining, or downgrading.

[0029] The early warning response module includes a thermal fault handling unit, a temperature fault management unit, and a cold fault archiving unit; The thermal fault handling unit is used to automatically trigger a high-level emergency response mechanism for pipeline areas that are determined to be thermal faults. The mechanism includes sending real-time audible and visual alarms to the operation and maintenance terminal, highlighting and flashing marks on the GIS map, automatically generating emergency inspection task work orders, and planning the optimal path to the fault point. The temperature fault management unit is used to add pipeline areas identified as having temperature faults to the system's priority inspection list, increase their task priority in subsequent periodic inspection plans, and generate potential risk investigation and recommendation reports. The cold fault archiving unit is used to automatically archive and store pipeline areas identified as cold faults, and generate historical fault analysis reports that include fault location, type, status time, and evolution process analysis, providing data support for optimizing long-term pipeline maintenance strategies.

[0030] The visualization module includes a GIS platform integration unit, a differentiated rendering unit, and an interactive analysis unit; The GIS platform integration unit is used to import fault information sets, including fault spatial location, fault type, fault level, and fault status time, into the geographic information system platform, and perform spatial registration and overlay with the underlying pipeline network base map. Specifically, this includes: Data format standardization: Receive the set of fault information from the decision output end, and convert the fault spatial location, fault type, fault level and fault status time fields into a standard spatial data format that can be recognized by the geographic information system platform; Coordinate System 1 and Transformation: The spatial coordinates of the fault information set are transformed from their original acquisition coordinate system to a system compatible with the pipeline network foundation. Figure 1 The mathematical relationship for coordinate transformation in a standardized engineering coordinate system can be expressed as follows: ; in, Represents latitude and longitude coordinates in the source coordinate system. This represents the transformed target plane rectangular coordinates. This represents the mapping function from the source geographic coordinate system to the target engineering coordinate system. This is a set of transformation parameters obtained in advance through control point calibration; Spatial registration: In the geographic information system platform, the fault point layer that has completed coordinate transformation is aligned with the corresponding pipeline element layer in the pipeline network base map to ensure that the fault point is located on the spatial location of its respective pipeline entity. Attribute association and layer overlay: Non-spatial attribute information in the fault information set, including fault type, fault level and fault status time, is associated with the corresponding spatial features of the fault point as an attribute table; then, the fault point layer is overlaid as a new thematic layer in the GIS view containing the pipeline network base map. The differentiated rendering unit is used to render and display different visual identification strategies on the GIS map according to the fault level: the first preset visual style is used to identify thermal faults, the second preset visual style is used to identify warm faults, and the third preset visual style is used to identify cold faults. The difference in visual styles is reflected in the shape, color, brightness and dynamic effect of the identification symbols. The interactive analysis unit provides a map interaction interface, allowing maintenance personnel to click and query fault icons displayed on the GIS map, filter regions, and perform multi-dimensional statistics. Responding to user operations, it dynamically displays detailed attribute information of the selected fault, associated multi-source feature evidence, and system-recommended handling measures.

[0031] The operation steps of this pipeline fault intelligent identification and location system based on UAV technology are as follows: Step 1: Collaborative Acquisition of Multi-Source Data The system starts operating via the data acquisition terminal. Equipped with multiple sensors, the drone automatically conducts flight inspections of the target pipeline according to a preset cruise path and time intervals. During flight, the drone simultaneously collects multi-source sensing data on the pipeline and its surrounding environment, including images, 3D point clouds, sound waves, specific gas concentrations, and temperature fields. Simultaneously, the flight control module generates observation sequence information for each acquisition event, containing a flight log sequence number, acquisition timestamp, and spatial coordinate sequence. All sensing data and observation sequence information are integrated and timestamped in the data encapsulation module, packaged into a unified sensing data packet, and sent to the data processing terminal in real time.

[0032] Step Two: Data Processing and Intelligent Analysis Upon receiving the sensing data packets, the data processing unit performs core analysis and diagnosis. First, the mapping construction module parses the data packets and, using the observation sequence identifier as an index, constructs a multi-dimensional mapping network connecting acquisition time, spatial location, environmental parameters, and sensor status. Next, the feature analysis module performs specialized analysis on various types of sensing data based on this mapping: the image analysis unit identifies color and texture anomalies and structural defects; the point cloud analysis unit calculates pipeline deformation and displacement; the acoustic signature analysis unit extracts abnormal acoustic features; the gas analysis unit detects excessive gas concentrations; and the thermal imaging analysis unit detects abnormal temperature fields. Each extracted key feature parameter is assigned its corresponding acquisition timestamp and spatial location label.

[0033] Subsequently, the intelligent association module performs deep fusion analysis on the aforementioned multi-source features with spatiotemporal labels. The data aggregation unit aggregates different features at the same spatial location. The pattern matching unit compares the aggregated feature set with a preset fault mode knowledge base, calculates the similarity between the comprehensive feature vector and the fault template, and determines the fault type corresponding to the location based on a preset matching confidence threshold. The time inversion unit traces the time series of the anomaly score of the feature parameters, identifies the time point when it first exceeds the static threshold, and reversely derives the state time of the fault.

[0034] Step 3: Fault Classification and Visual Response: The decision output terminal receives fault type and fault status time information and executes a decision. The fault classification module calculates the interval between the fault status time and the current time, and automatically classifies faults into "hot faults" requiring urgent handling, "warm faults" requiring priority attention, and "cold faults" that can be archived and managed by comparison with preset time thresholds. It can also dynamically adjust the weights based on historical data and environmental stress. According to the classified level, the early warning response module automatically triggers differentiated responses: real-time alarms are issued for hot faults and emergency work orders are generated; the priority of subsequent inspections is increased for warm faults; and cold faults are archived and analysis reports are generated.

[0035] Meanwhile, the visualization module imports fault information, including fault location, type, level, and status time, into the geographic information system. The GIS platform integration unit performs coordinate transformation, spatial registration, and overlay with the pipeline base map. The differentiated rendering unit then uses icons with different colors, shapes, and dynamic effects to prominently identify faults on the electronic map based on their level. Maintenance personnel can use the interactive analysis unit to query fault details, related multi-source evidence, and system handling suggestions, thereby achieving intelligent identification, location, and tiered response to pipeline faults.

[0036] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0037] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A pipeline fault intelligent identification and location system based on UAV technology, characterized in that: The system includes a data acquisition terminal, a data processing terminal, and a decision output terminal; The data acquisition terminal is used to collect multi-source sensing data and its observation sequence information of the pipeline by a drone equipped with multiple sensors according to a preset path and interval, and send it to the data processing terminal. The multi-source sensing data includes image, point cloud, sound wave, gas concentration and temperature field data, and the observation sequence information includes flight log sequence number, collection timestamp and spatial coordinate sequence. The data processing terminal is used to receive the multi-source sensing data and observation sequence information, construct a multi-dimensional mapping relationship between the observation sequence information and the acquisition time, spatial location and environmental parameters, and based on this mapping relationship, extract key feature parameters with spatiotemporal labels from the multi-source sensing data, and perform cross-scale correlation analysis on the feature parameters to jointly determine the fault type and reversely deduce the fault state time. The decision output terminal is used to receive the fault type and fault status time. By comparing the time interval between the fault status time and the current time with a preset threshold, the fault level is determined as a hot fault, a warm fault, or a cold fault. Different fault pipeline areas of different levels are differentiated and spatially located in the geographic information system.

2. The intelligent pipeline fault identification and location system based on UAV technology according to claim 1, characterized in that: The data acquisition terminal includes a flight control module, a sensor array module, and a data encapsulation module; The flight control module includes a path planning unit, a state synchronization unit, and a sequence generation unit; The path planning unit is used to plan the preset cruise path of the UAV based on the geographical information of the pipeline to be monitored, and to set the time interval for data collection. The state synchronization unit is used to synchronize and record environmental parameters, UAV platform flight status parameters and sensor working status parameters in real time at each acquisition moment during UAV flight. The environmental parameters include wind speed, light intensity and ambient temperature, and the platform flight status parameters include flight altitude, flight speed and attitude angle. The sequence generation unit is used to generate a unique observation sequence identifier for each data acquisition event. The observation sequence identifier includes the track point number, image frame sequence number, laser point cloud timestamp sequence number, and acoustic sampling sequence number.

3. The intelligent pipeline fault identification and location system based on UAV technology according to claim 2, characterized in that: The sensor array module includes an optical sensing unit, a point cloud sensing unit, an acoustic wave sensing unit, a gas sensing unit, and a temperature sensing unit. The optical sensing unit is used to acquire optical and infrared image data of the pipeline and its surrounding environment through a high-resolution optical camera and infrared thermal imager mounted on the UAV. The point cloud sensing unit is used to collect three-dimensional point cloud data of pipelines and their ancillary facilities through the lidar equipment carried by the UAV. The acoustic wave sensing unit is used to collect acoustic wave signals and vibration signal data generated during pipeline operation through acoustic wave sensors and acoustic emission sensors carried by the UAV. The gas sensing unit is used to collect concentration data of specific gas components in the air around the pipeline using a gas sensor carried by the UAV. The temperature sensing unit is used to collect temperature field distribution data on the pipeline surface through infrared temperature measurement equipment and temperature sensor array carried by the UAV.

4. The intelligent pipeline fault identification and location system based on UAV technology according to claim 2, characterized in that: The data encapsulation module includes a data receiving unit, a timestamp binding unit, and a data packet generation unit; The data receiving unit is used to receive multi-source sensing data from the sensor array module in real time, as well as observation sequence identifiers, environmental parameters, and UAV platform status parameters from the flight control module. The timestamp binding unit is used to bind the corresponding timestamp during the acquisition process to each set of multi-source sensing data and its corresponding observation sequence identifier received; The data packet generation unit is used to integrate and encapsulate multi-source sensing data with timestamps, observation sequence identifiers, corresponding UAV spatial coordinates, environmental parameters, and platform status parameters into a sensing data packet with a unified sequence identifier, and send it to the data processing terminal.

5. The intelligent pipeline fault identification and location system based on UAV technology according to claim 1, characterized in that: The data processing terminal includes a mapping construction module, a feature parsing module, and an intelligent association module; The mapping construction module includes a sequence parsing unit, a spatiotemporal association unit, and a context binding unit; The sequence parsing unit is used to parse the received sensing data packets and extract the observation sequence identifier, collection timestamp and UAV spatial coordinates. The spatiotemporal correlation unit is used to construct a forward mapping chain from the sequence identifier to the acquisition timestamp and then to the spatial coordinates, using the observation sequence identifier as an index, and to establish a corresponding reverse query index. The context binding unit is used to further bind the acquisition timestamp with environmental parameters, platform flight status parameters, and sensor operating status parameters to form a multi-dimensional mapping relationship network for subsequent comprehensive analysis.

6. The intelligent pipeline fault identification and location system based on UAV technology according to claim 5, characterized in that: The feature parsing module includes an image analysis unit, a point cloud analysis unit, a voiceprint analysis unit, a gas analysis unit, and a thermal imaging analysis unit; The image analysis unit is used to perform color analysis, texture analysis and contour recognition on optical and infrared images based on the multi-dimensional mapping relationship, extract color abnormality areas, texture abnormality areas, structural cracks, leakage reflection and surface corrosion feature parameters in the images, and assign a corresponding acquisition timestamp and spatial location label to each feature parameter. The point cloud analysis unit is used to perform registration and comparison analysis on the three-dimensional point cloud data based on the multi-dimensional mapping relationship, calculate the deformation, displacement, settlement and geometric anomaly index of the pipeline, and assign a corresponding collection timestamp and spatial location label to each deformation index. The acoustic signature analysis unit is used to perform time-frequency domain transformation and feature extraction on the acoustic signal based on the multi-dimensional mapping relationship, identify abnormal frequency components, energy mutations and acoustic emission features of specific modes in the signal, and assign a corresponding acquisition timestamp and spatial location label to each acoustic feature. The gas analysis unit is used to perform component analysis and threshold comparison on the concentration data collected by the gas sensor based on the multi-dimensional mapping relationship, identify the concentration exceedance and leakage characteristics of specific gas components, and assign a corresponding collection timestamp and spatial location label to each gas anomaly characteristic. The thermal imaging analysis unit is used to perform temperature field analysis and region segmentation on the thermal imaging data acquired by the infrared thermal imager based on the multi-dimensional mapping relationship, extract temperature anomaly regions, temperature gradient anomalies and hot spot features, and assign a corresponding acquisition timestamp and spatial location label to each thermal feature.

7. The intelligent pipeline fault identification and location system based on UAV technology according to claim 5, characterized in that: The intelligent association module includes a data aggregation unit, a pattern matching unit, and a time inversion unit; The data aggregation unit is used to aggregate and align key feature parameters with spatiotemporal labels extracted from different sensors at different times in the same pipeline area, with spatial location as the core, to form a feature parameter set for that location. The pattern matching unit is used to match and analyze the set of feature parameters with a preset fault mode knowledge base. The fault mode knowledge base stores multi-source feature combination templates corresponding to various fault types. Through matching, the fault type corresponding to the pipeline area is identified. The fault types include minor leaks, gas leaks, material aging, external corrosion, structural cracks, and mechanical damage. The time inversion unit is used to perform reverse derivation of the fault occurrence time based on the multi-dimensional mapping relationship and the feature parameter set. Specifically, it traverses the anomaly scores of each feature parameter in the feature parameter set in sequence on the time axis, identifies the time node when the anomaly score first exceeds the preset threshold, and determines the time node as the fault state time of the pipeline area.

8. The intelligent pipeline fault identification and location system based on UAV technology according to claim 1, characterized in that: The decision output terminal includes a level determination module, an early warning response module, and a visualization presentation module; The grade determination module includes an interval calculation unit, a threshold comparison unit, and a dynamic weighting unit; The interval calculation unit is used to calculate the time interval between the fault status time output by the data processing terminal and the current system time; The threshold comparison unit is used to compare the time interval with a preset first threshold and a second threshold, and classify the fault level according to the comparison result; The dynamic weighting unit is used to dynamically correct and adjust the initially determined fault level based on threshold comparison, combined with historical fault frequency data and environmental stress condition data of the faulty pipeline area.

9. A pipeline fault intelligent identification and location system based on UAV technology according to claim 8, characterized in that: The early warning response module includes a thermal fault handling unit, a temperature fault management unit, and a cold fault archiving unit. The thermal fault handling unit is used to automatically trigger a high-level emergency response mechanism for pipeline areas that are determined to be thermal faults. The mechanism includes sending real-time audible and visual alarms to the operation and maintenance terminal, highlighting and flashing marks on the GIS map, automatically generating emergency inspection task work orders, and planning the optimal path to the fault point. The temperature fault management unit is used to include pipeline areas identified as having temperature faults in the system's priority inspection list, increase their task priority in subsequent periodic inspection plans, and generate potential risk investigation and suggestion reports. The cold fault archiving unit is used to automatically archive and store pipeline areas identified as cold faults, and generate historical fault analysis reports containing fault location, type, status time, and evolution process analysis, providing data support for optimizing long-term pipeline maintenance strategies.

10. A pipeline fault intelligent identification and location system based on UAV technology according to claim 8, characterized in that: The visualization module includes a GIS platform integration unit, a differentiated rendering unit, and an interactive analysis unit; The GIS platform integration unit is used to import the fault information set, which includes the fault spatial location, fault type, fault level and fault status time, into the geographic information system platform, and perform spatial registration and overlay with the underlying pipeline network base map. The differentiated rendering unit is used to render and display different visual identification strategies on the GIS map according to the fault level: a first preset visual style is used to identify thermal faults, a second preset visual style is used to identify warm faults, and a third preset visual style is used to identify cold faults. The difference in the visual style is reflected in the shape, color, brightness and dynamic effect of the identification symbols. The interactive analysis unit provides a map interaction interface, enabling maintenance personnel to click and query fault icons displayed on the GIS map, filter regions, and perform multi-dimensional statistics. In response to user operations, it dynamically displays detailed attribute information of the selected fault, associated multi-source feature evidence, and system-recommended handling measures.

Images