Pipeline detection method and system based on unmanned aerial vehicle

Abstract

The invention discloses a pipeline detection method and system based on an unmanned aerial vehicle, relates to the technical field of unmanned aerial vehicle detection, and comprises the steps of selecting a reasonable unmanned aerial vehicle collection point by considering the condition of a to-be-detected pipeline, and providing a reliable basis for subsequent flight plan planning and pipeline detection of the unmanned aerial vehicle. An initial flight plan is formulated by considering the environment condition of the position where the pipeline is located and the requirement for flight stability, and the adaptability of flight of the unmanned aerial vehicle is guaranteed. And judging whether second flight is needed or not based on the quality of the first detection information, judging whether the first flight is enough to support the pipeline detection quality requirement or not according to the quality condition of the detection information, and outputting a pipeline detection result by considering complementarity and intersectionality between the first detection information and the second detection information. The adaptability and accuracy of pipeline detection are improved, the adaptability of the flight condition of the unmanned aerial vehicle and the pipeline environment is ensured, and the requirement of the complex environment where the pipeline is located for detection is met.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle (UAV) detection, and particularly to a pipeline detection method and system based on UAVs. Background Art

[0002] In traditional pipeline inspection, manual inspection faces problems such as complex terrain, low efficiency, high cost, and high danger, making it difficult to meet the requirements of large-scale pipeline safety monitoring. With the rapid development of UAV technology and artificial intelligence, pipeline detection solutions based on UAVs have emerged. This solution integrates technologies such as UAV flight platforms, high-definition sensors, AI image recognition, edge computing, and 3D modeling, enabling efficient coverage of complex terrains and real-time collection of pipeline and surrounding environment data. By carrying devices such as high-definition cameras, lidar, and infrared thermal imagers, UAVs can accurately identify defects such as cracks, corrosion, and leaks on the pipeline surface, and use AI algorithms to analyze the data in real time. At the same time, 3D real-scene modeling technology can convert image data into high-precision 3D models to assist in defect location and risk assessment. This solution breaks through the geographical limitations of traditional inspections, significantly improving the detection efficiency and accuracy, and providing an intelligent and digital solution for the safe operation of pipelines.

[0003] In the prior art, the impact of UAV flight conditions on pipeline detection information is not considered, that is, the adaptability between UAV flight conditions and the pipeline detection environment is not taken into account, resulting in poor adaptability and accuracy of pipeline detection and unable to meet the requirements of complex and changing pipeline detection.

[0004] Therefore, how to improve the adaptability and accuracy of pipeline detection is a technical problem to be solved at present. Summary of the Invention

[0005] The purpose of the present invention is to solve the problem in the prior art that due to the lack of consideration of the adaptability between UAV flight conditions and the pipeline detection environment, the adaptability and accuracy of pipeline detection are poor, and a pipeline detection method based on UAVs is proposed, which includes: Obtain the pipeline information, pipeline position environment information, pipeline detection task, and UAV-borne sensor information of the pipeline to be measured, set the acquisition section of the pipeline according to the pipeline information and pipeline detection task of the pipeline to be measured, and formulate the first flight plan based on the pipeline position environment information and the acquisition section of the pipeline; The UAV conducts flight according to the first flight plan, collects the first detection information of the pipeline, counts the flight information deviation between the first actual flight and the first flight plan, evaluates the quality of the first detection information, and determines whether a second flight is required based on the quality of the first detection information; If a second flight is required, optimize the second flight plan according to the flight information deviation, and the UAV flies according to the second flight plan to collect the second detection information of the pipeline; Output the detection result of the pipeline by integrating the first detection information and the second detection information of the pipeline.

[0006] In some embodiments of the present application, the acquisition section of the pipeline is set according to the pipeline information and the pipeline detection task of the pipeline to be measured, including, The pipeline information includes the pipeline length, pipeline orientation, pipeline material, and the medium transported in the pipeline; Define a simple pipeline scenario and a complex pipeline scenario according to the pipeline length and pipeline orientation. For the simple pipeline scenario, the entire pipeline to be measured is used as the acquisition section of the pipeline to be measured; For the complex pipeline scenario, establish a pipeline model of the pipeline to be measured according to the pipeline length, pipeline orientation, pipeline material, and the medium transported in the pipeline. Collect all risk elements through the pipeline detection task, and mark the high-risk positions at the corresponding positions of the pipeline model of the pipeline to be measured by integrating all risk elements. The high-risk positions are used as the acquisition section of the pipeline to be measured.

[0007] In some embodiments of the present application, define a simple pipeline scenario and a complex pipeline scenario according to the pipeline length and pipeline orientation, including, Extract multiple key features on the pipeline orientation, divide the pipeline into multiple sections according to the pipeline length, count the key feature density under each section to obtain the average key feature density, define the scene complexity based on the pipeline length, multiple key features, and the average key feature density, and distinguish the simple pipeline scenario and the complex pipeline scenario through the scene complexity.

[0008] In some embodiments of the present application, formulate the first flight plan based on the pipeline position environment information and the acquisition section of the pipeline, including, Import the pipeline information and the pipeline position environment information into the GIS software, and overlay the obstacle distribution map in the pipeline position environment. Mark the position of the acquisition section of the pipeline in the GIS software. Using multiple acquisition section positions as waypoints, generate multiple candidate paths through the path planning algorithm, and each candidate path passes through all waypoints; Split each candidate path into multiple path units, evaluate the path cost and flight risk cost of each candidate path based on the path units, generate a comprehensive cost by combining the path cost and the flight risk cost, and draw a comprehensive cost change curve that changes with the path units; Distinguish the stable part curve and the fluctuating part curve on the comprehensive cost change curve, calculate the relative cost of the stable part curve and the fluctuating part curve respectively through the change of the curve slope, so as to determine the relative comprehensive cost, and select a candidate path as the target path by virtue of the relative comprehensive cost; Configure the flight information of the drone based on the target path.

[0009] In some embodiments of the present application, configuring the flight information of the drone based on the target path includes, Dividing the target path into a flight path and a collection path according to the position of the collection section of the pipeline; For the flight path, analyze the requirements of the environment for flight stability through GIS software, and configure the flight information of the drone in combination with the requirements of the flight mission; For the collection path, analyze the requirements of the pipeline for flight stability and the requirements of the environment for flight stability through GIS software, and configure the flight information of the drone based on the requirements of the pipeline for flight stability, the requirements of the environment for flight stability, and the requirements of the flight mission.

[0010] In some embodiments of the present application, statistically analyze the deviation of flight information between the first actual flight and the first flight plan, including, Statistically analyze the deviation of the flight information of the flight path and the collection path respectively, obtain the first deviation and the second deviation, draw the respective change curves of the first deviation and the second deviation changing with position, identify the respective deviation points on the respective change curves of the first deviation and the second deviation, and analyze the correlation between the two types of deviation points.

[0011] In some embodiments of the present application, optimize the second flight plan according to the flight information deviation, including, Spatially align the flight information deviation and the quality of the first detection information through the pipeline position, and optimize to obtain the second flight plan based on the correlation between the deviation points.

[0012] In some embodiments of the present application, output the detection result of the pipeline by integrating the first detection information and the second detection information of the pipeline, including, Determine the intersection information part and the difference set information part between the first detection information and the second detection information according to the position of the pipeline; Fuse the intersection information part of the first detection information and the second detection information to obtain the fusion information; Output the detection result of the pipeline based on the difference set information part and the fusion information.

[0013] Correspondingly, the present application also provides a pipeline detection system based on a drone, including, The first module is used to obtain the pipeline information, pipeline position environment information, pipeline detection task, and drone-borne sensor information of the pipeline to be detected, set the collection section of the pipeline according to the pipeline information and pipeline detection task of the pipeline to be detected, and formulate the first flight plan based on the pipeline position environment information and the collection section of the pipeline; The second module is used for the drone to fly according to the first flight plan, collect the first detection information of the pipeline, count the flight information deviation between the first actual flight and the first flight plan, evaluate the quality of the first detection information, and judge whether a second flight is needed based on the quality of the first detection information; The third module is used for, if a second flight is needed, optimizing the second flight plan according to the flight information deviation, and the drone flies according to the second flight plan to collect the second detection information of the pipeline; The fourth module is used for outputting the detection result of the pipeline by integrating the first detection information and the second detection information of the pipeline.

[0014] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. Set the acquisition section of the pipeline according to the pipeline information and pipeline detection task of the pipeline to be measured, select reasonable drone acquisition points considering the actual situation of the pipeline to be measured, and provide a reliable basis for the subsequent flight plan planning of the drone and pipeline detection. Formulate the first flight plan based on the pipeline position environment information and the acquisition section of the pipeline, and formulate the initial flight plan considering the position environment situation of the pipeline and the demand for flight stability, ensuring the adaptability of the drone flight.

[0015] 2. Judge whether a second flight is needed based on the quality of the first detection information, and judge whether the first flight is sufficient to support the pipeline detection quality requirements according to the quality of the detection information. When it is difficult to support, conduct a second flight. Optimize the second flight plan according to the flight information deviation to make up for the loopholes in the first flight, ensuring the reliability and adaptability of the second flight. Output the detection result of the pipeline by integrating the first detection information and the second detection information of the pipeline, and consider the complementarity and intersection between the first detection information and the second detection information to output the detection result of the pipeline, improving the adaptability and accuracy of pipeline detection, ensuring the adaptability of the drone flight situation to the pipeline environment, and meeting the detection requirements of the complex environment where the pipeline is located. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 is a schematic flowchart of a pipeline detection method based on a drone proposed by the present invention; Figure 2 is a schematic structural diagram of a pipeline detection system based on a drone proposed by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0017] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments.

[0018] Refer toFigure 1 , a pipeline detection method based on an unmanned aerial vehicle (UAV), comprising the following steps: Step S101: Obtain pipeline information, pipeline location environment information, pipeline detection tasks, and UAV-borne sensor information of the pipeline to be measured. Set the acquisition section of the pipeline according to the pipeline information and pipeline detection tasks of the pipeline to be measured, and formulate a first flight plan based on the pipeline location environment information and the acquisition section of the pipeline.

[0019] In this embodiment, as an important part of systems such as energy transmission and urban water supply and drainage, the safety and stability of pipelines are crucial. To timely detect potential problems existing in pipelines and ensure the normal operation of the pipeline system, using a UAV for pipeline detection has become an efficient and convenient method. Obtaining detailed information of the pipeline to be measured is the basis for formulating a flight plan. Such information includes the type of the pipeline (such as oil and gas pipelines, natural gas pipelines, water supply and drainage pipelines, etc.), the material of the pipeline, the diameter and length of the pipeline, etc. Different types of pipelines may require different sensors and detection methods during detection. The material and diameter of the pipeline will affect the detection sensitivity of the sensor to pipeline defects, while the burial depth of the pipeline will affect the flight altitude of the UAV and the penetration ability of the detection equipment. Understanding the environmental information of the pipeline location is crucial for formulating a flight plan. This includes the terrain and landform around the pipeline (such as mountains, plains, cities, etc.), meteorological conditions (such as wind speed, wind direction, temperature, humidity, etc.), and the distribution of surrounding buildings and obstacles. The terrain and landform will affect the flight route of the UAV and the selection of take-off and landing points. Meteorological conditions will affect the flight safety of the UAV and the accuracy of detection data. Surrounding buildings and obstacles require the UAV to avoid during flight to ensure flight safety. Defining the pipeline detection task is the key to determining the detection objectives and requirements. The detection task may include detecting whether there are defects such as leaks, corrosion, deformation, and rupture in the pipeline, and determining the location, size, and severity of the defects. Different detection tasks require different sensors and detection methods, and have different requirements for the accuracy and integrity of detection data. Understanding the performance and characteristics of the sensors carried by the UAV is an important basis for formulating a flight plan. The sensors may include high-resolution cameras, thermal imagers, lidar, gas sensors, etc. Different sensors have different detection principles and applicable ranges. For example, a high-resolution camera can be used to detect deformation and damage on the pipeline surface, a thermal imager can be used to detect pipeline leaks and abnormal temperature distributions, lidar can be used to obtain three-dimensional terrain information around the pipeline, and a gas sensor can be used to detect whether there is leakage of combustible or toxic gases around the pipeline.

[0020] In some embodiments of the present application, setting the acquisition section of the pipeline according to the pipeline information and pipeline detection tasks of the pipeline to be measured includes: The pipeline information includes pipeline length, pipeline orientation, pipeline material, and the medium transported inside the pipeline; Define simple pipeline scenarios and complex pipeline scenarios based on the pipeline length and pipeline orientation. For simple pipeline scenarios, the entire pipeline to be measured is taken as the acquisition section of the pipeline to be measured. For complex pipeline scenarios, establish a pipeline model of the pipeline to be measured according to the pipeline length, pipeline orientation, pipeline material, and the medium transported inside the pipeline. Collect all risk elements through pipeline inspection tasks, and mark the high-risk positions at the corresponding positions of the pipeline model of the pipeline to be measured by integrating all risk elements. Take the high-risk positions as the acquisition section of the pipeline to be measured.

[0021] In this embodiment, the following information is extracted from pipeline design drawings or databases: Pipeline length (L): in meters (m), used for section division.

[0022] Pipeline orientation: Describe the extension direction of the pipeline through geographical coordinates or mileage coordinates.

[0023] Pipeline material (such as steel pipe, plastic pipe): Affects the corrosion rate and defect type.

[0024] Medium transported inside the pipeline (such as oil, natural gas, water): Affects the leakage risk and inspection focus.

[0025] Extract key features such as bends, branches, undulations, slope changes, intersections, etc. from the pipeline orientation. Divide simple scenarios and complex scenarios. For simple scenarios, the entire pipeline can be used as the acquisition section. For complex scenarios, in order to balance the detection efficiency and accuracy, it is necessary to select part of the pipeline as the acquisition section.

[0026] Risk element collection: According to pipeline inspection tasks, collect possible risk elements (such as stress concentration at bends, leakage risk at branches, corrosion-sensitive areas of materials, etc.).

[0027] Mark high-risk positions: Map the risk elements to the pipeline model and mark the high-risk positions (such as bends, branches, material aging areas, etc.).

[0028] Set the acquisition section: Take the high-risk positions and their surrounding areas as the acquisition section, and increase the acquisition density (such as reducing the flight altitude, increasing the acquisition angle).

[0029] In some embodiments of the present application, define simple pipeline scenarios and complex pipeline scenarios according to the pipeline length and pipeline orientation, including, Extract multiple key features on the pipeline orientation, divide the pipeline into multiple sections according to the pipeline length, count the key feature density under each section, obtain the average key feature density, define the scenario complexity based on the pipeline length, multiple key features, and the average key feature density, and distinguish simple pipeline scenarios and complex pipeline scenarios through the scenario complexity.

[0030] In this embodiment, the scene complexity calculation formula is as follows: Wherein, is the scene complexity, is the pipeline length coefficient, is the number of key features, is the th combination weight of the key feature, is the th complexity of the key feature. The key features may include bends, branches, undulations, etc., and the complexity may be the angle of the bend, the level of branching, the degree of undulation, etc., is the th average density of the key feature (average frequency of occurrence in the section), is the th first constant of the key feature, represents the correction of the sum of the complexities of the key features by the average density of the key features.

[0032] Key feature extraction: Extract key features such as bends, branches, undulations, slope changes, intersections, etc. from the pipeline route.

[0033] Section division: Divide the pipeline into multiple continuous sections according to the pipeline length (for example, every 100 meters is a section).

[0034] Feature density calculation: Count the number of key features in each section and calculate the key feature density (number of features / section length).

[0035] Quantify the scene complexity through the key feature density and the pipeline length, avoid subjective judgment, and improve the scientificity of the acquisition section setting. In complex scenes, high-risk positions (such as bends, branches, material aging areas) are high-incidence areas of defects and need to be focused on for acquisition.

[0036] In some embodiments of the present application, a first flight plan is formulated based on the pipeline position environment information and the acquisition section of the pipeline, including, Import the pipeline information and the pipeline position environment information into the GIS software, and overlay the obstacle distribution map in the pipeline position environment. Mark the acquisition section positions of the pipeline in the GIS software. Using the multiple acquisition section positions as waypoints, generate multiple candidate paths through the path planning algorithm, and each candidate path passes through all waypoints; Split each candidate path into multiple path units, evaluate the path cost and the flight risk cost of each candidate path based on the path units, generate a comprehensive cost by combining the path cost and the flight risk cost, and draw a comprehensive cost change curve that changes with the path units; Distinguish the stable part curve and the fluctuating part curve on the comprehensive cost change curve, calculate the relative costs of the stable part curve and the fluctuating part curve respectively through the change of the curve slope, so as to determine the relative comprehensive cost, and select a candidate path as the target path based on the relative comprehensive cost; Configure the flight information of the UAV based on the target path.

[0037] In this embodiment, the path planning algorithm: uses the A* algorithm, RRT (Rapidly-exploring Random Tree) or ant colony algorithm to generate multiple candidate paths to ensure that each path passes through all passing points.

[0038] Obstacle superposition: Superpose the obstacle distribution map (such as buildings, trees, high-voltage lines, etc.) in GIS.

[0039] Collection section annotation: Mark the collection section positions of the pipeline (such as high-risk positions, complex field sections) as passing points.

[0040] Path unit splitting: Split each candidate path into multiple path units (such as every 10 meters as a unit).

[0041] Path cost calculation: For example, energy consumption cost: Calculate the energy consumption (unit: joules / meter) according to the flight speed, altitude and load of the UAV.

[0042] Risk cost calculation: Obstacle distance cost: The closer to the obstacle, the higher the cost (for example, when the distance < 5 meters, the cost = 100; when the distance is 5 - 10 meters, the cost = 50).

[0043] Because the two types of costs will change at different positions, the cost evaluation is based on path units. Combine the path cost and the flight risk cost to generate a comprehensive cost (which can be achieved by weighted summation, etc.).

[0044] Input data: The comprehensive cost (path cost + flight risk cost) based on the path unit splitting.

[0045] Curve drawing: Use the path unit as the horizontal axis (X-axis) and the comprehensive cost as the vertical axis (Y-axis) to draw a continuous comprehensive cost change curve.

[0046] The stable part curve is the curve where the comprehensive cost appears more frequently, and the fluctuating part curve is the curve where the comprehensive cost appears less frequently. Calculate the average value of the slope changes of the stable part curve and the fluctuating part curve respectively, and weight according to the occurrence frequency of each comprehensive cost under the comprehensive cost interval of the stable part curve and the fluctuating part curve to obtain the representative values of the comprehensive costs of the two curves respectively. Based on the representative values of the comprehensive costs and the average value of the slope changes, obtain the relative comprehensive cost. The calculation formula is as follows: Among them, is the th candidate path (corresponding comprehensive cost change curve), , are the cost weights of the stable part curve and the fluctuating part curve respectively. Generally speaking, the cost weight of the stable part curve is higher than that of the fluctuating part curve. , are the respective comprehensive cost representative values of the stable part curve and the fluctuating part curve of the th candidate path, , are the respective average slope change values of the stable part curve and the fluctuating part curve of the th candidate path, , are the second constant and the third constant corresponding to the th candidate path respectively, and represent the correction of the average slope change value to the comprehensive cost representative value. The more stable the change, the greater the credibility. The second constant and the third constant are used to balance the magnitude of the correction function.

[0048] In some embodiments of the present application, based on the target path, the flight information of the drone is configured, including dividing the target path into a flight path and a collection path according to the position of the acquisition section of the pipeline; For the flight path, analyze the environmental requirements for flight stability through GIS software and configure the flight information of the drone in combination with the flight mission requirements; For the collection path, analyze the pipeline's requirements for flight stability and the environmental requirements for flight stability through GIS software, and configure the flight information of the drone based on the pipeline's requirements for flight stability, the environmental requirements for flight stability, and the flight mission requirements.

[0049] In this embodiment, the flight path: in the target path, the drone only needs to fly to the transition section between the acquisition sections (without performing acquisition tasks).

[0050] The collection path: the section of the target path where the drone needs to collect data on the pipeline (such as shooting, detection).

[0051] Environmental stability requirements: Analyze the terrain (such as plains, mountains), meteorology (such as wind speed, precipitation), and obstacle distribution (such as buildings, high-voltage lines) of the flight path through GIS software.

[0052] Flight mission requirements: Such as path requirements, time requirements, etc.

[0053] Configuration parameters: Flight altitude and attitude: Adjust according to the height of obstacles (e.g., when avoiding high-voltage lines, the flight altitude > 20 meters).

[0054] Flight speed: Adjust according to the wind speed (e.g., when the wind speed < 3 m / s, the speed = 8 m / s; when the wind speed > 5 m / s, the speed = 5 m / s).

[0055] Collision avoidance strategy: Enable the radar or visual collision avoidance system.

[0056] Pipeline stability requirements: Pipeline type (e.g., buried, overhead), acquisition accuracy requirements (e.g., low speed flight is required for high-precision).

[0057] Environmental stability requirements: Same as the flight path analysis.

[0058] Configuration parameters: Flight altitude: Adjust according to the pipeline position (e.g., for overhead pipelines, the flight altitude = 10 meters).

[0059] Flight speed: Adjust according to the acquisition accuracy (e.g., when performing high-precision acquisition, the speed = 3 m / s).

[0060] Shooting angle: Adjust according to the pipeline direction (e.g., vertical shooting or inclined shooting).

[0061] Collision avoidance strategy: Enable the collision avoidance system to ensure the safety of the acquisition equipment.

[0062] In step S102, the UAV performs a flight according to the first flight plan, acquires the first detection information of the pipeline, counts the flight information deviation between the first actual flight and the first flight plan, evaluates the quality of the first detection information, and determines whether a second flight is required based on the quality of the first detection information.

[0063] In this embodiment, the flight task is started: The UAV executes the flight task according to the pre-generated first flight plan (including the target path, flight altitude, speed, acquisition parameters, etc.).

[0064] Data acquisition: In the acquisition path segment, the UAV acquires the first detection information of the pipeline (such as surface defects, temperature distribution, deformation data) through the sensors carried (such as high-definition cameras, infrared thermal imagers, lidar).

[0065] Deviation type: Path deviation: Lateral / longitudinal deviation of the actual flight trajectory from the target path (unit: meter).

[0066] Position deviation: Position deviation of the actual reached waypoint from the planned waypoint (unit: meter).

[0067] Attitude deviation: The deviation between the actual flight attitude (pitch angle, roll angle, yaw angle) and the planned attitude (unit: degree).

[0068] Speed deviation: The deviation between the actual flight speed and the planned speed (unit: m / s).

[0069] Deviation statistical method: By comparing the flight log (such as GPS data, IMU data) recorded by the UAV flight control system with the planned parameters, calculate the mean and maximum values of each deviation value.

[0070] First detection information quality assessment (enumerated, not exhaustive) Evaluation indicators: Data integrity: Whether the collected pipeline data covers all planned collection segments (such as the proportion of the missing segment length).

[0071] Data clarity: Whether the resolution of the image or point cloud data meets the detection requirements (such as the proportion of blurred images).

[0072] Judgment rules: Impact of flight deviation: If the path deviation or attitude deviation causes data loss or blurring (such as the proportion of blurred images > 5%), a second flight is required.

[0073] Data quality threshold: If the data integrity < 90% or the data accuracy (false detection rate + missed detection rate) > 5%, a second flight is required.

[0074] In some embodiments of the present application, the flight information deviation between the first actual flight and the first flight plan is statistically analyzed, including Statistically analyze the deviations of the flight information of the flight path and the acquisition path respectively, obtain the first deviation and the second deviation, draw the respective change curves of the first deviation and the second deviation with the change of position, identify the respective deviation points on the respective change curves of the first deviation and the second deviation, and analyze the correlation between the two types of deviation points.

[0075] In this embodiment, deviation types: Flight path deviation (first deviation): Statistically analyze the deviation between the actual flight parameters (such as path, position, attitude, speed) of the flight path segment and the planned parameters.

[0076] Acquisition path deviation (second deviation): Statistically analyze the deviation between the actual flight parameters of the acquisition path segment and the planned parameters.

[0077] Statistical method: Export the flight log from the UAV flight control system and extract the actual flight data (GPS coordinates, IMU attitude, speed, etc.).

[0078] Align the actual data with the planned data by time or location, and calculate the deviation value (e.g., actual position - planned position).

[0079] Curve drawing rules: Use the path position (such as waypoint number or cumulative distance) as the horizontal axis (X-axis) and the deviation value as the vertical axis (Y-axis).

[0080] Draw the flight path deviation curve and the acquisition path deviation curve respectively.

[0081] Recognition rules: Definition of deviation points: Points where the deviation value exceeds the threshold (e.g., path deviation > 1.0 m, attitude deviation > 2.0°).

[0082] Recognition method: Mark the deviation points through a sliding window or peak detection algorithm.

[0083] Analysis of the correlation between two types of deviation points Analysis method: Cause correlation: Analyze the common causes of deviation points (e.g., sudden wind speed change, complex terrain leading to unstable flight attitude).

[0084] Causal association, the first type of deviation causes the second type of deviation: To determine whether the flight path deviation points (the first type of deviation points) cause the acquisition path deviation points (the second type of deviation points), it is necessary to start from the following aspects: Spatial correlation: Whether the two types of deviation points appear in close positions or the same flight section.

[0085] Temporal correlation: Whether the two types of deviation points occur continuously in time (e.g., the acquisition path deviation appears immediately after the flight path deviation).

[0086] Causal logic: Whether the first type of deviation points directly or indirectly cause the second type of deviation points through physical mechanisms (such as attitude out of control, speed mutation).

[0087] Spatial correlation analysis Method: Mark the flight path deviation points and the acquisition path deviation points on the flight trajectory map, and check whether they overlap or are adjacent.

[0088] Temporal correlation analysis Method: According to the flight log timestamps, check whether the two types of deviation points occur continuously within a short period of time.

[0089] Causal logic analysis Possible mechanisms: Path deviation → Attitude deviation: Flight path deviation (such as deviation from the planned route) may cause the UAV to adjust its attitude to correct the trajectory, thereby triggering attitude deviation.

[0090] Path deviation → Speed deviation → Attitude deviation: Path deviation may trigger speed adjustment (such as decelerating to correct the trajectory), and the sudden change in speed causes attitude instability.

[0091] Superposition of environmental factors: Path deviation may be caused by environmental interference (such as sudden change in wind speed), and the same environmental interference simultaneously causes deviation in the acquisition path.

[0092] In step S103, if a second flight is required, optimize the second flight plan according to the flight information deviation, and the UAV performs flight according to the second flight plan to collect the second detection information of the pipeline.

[0093] In some embodiments of the present application, optimizing the second flight plan according to the flight information deviation includes, Spatially align the flight information deviation and the quality of the first detection information through the pipeline position, and optimize to obtain the second flight plan based on the correlation between the deviation points.

[0094] In this embodiment, pipeline position mapping: Based on the actual position of the pipeline (such as mileage stake number, GPS coordinates), map the flight information deviation (path deviation, attitude deviation, etc.) and the quality of the first detection information (data integrity, clarity, accuracy) to the same spatial coordinate system. Generate a "deviation-quality" joint distribution map to clarify the spatial correspondence between deviation and quality problems.

[0095] Optimization strategy: Path optimization: Increase the density of waypoints in the high-incidence section of deviation points (such as mileage stake number 2km - 3km), reduce the flight interval, and improve the trajectory control accuracy.

[0096] Adjust the flight altitude or speed to reduce environmental interference (such as reducing the altitude to avoid wind interference from high-rise buildings).

[0097] Attitude control optimization: Enable a more stringent attitude stabilization algorithm (such as PID control parameter adjustment) in the high-incidence section of attitude deviation (such as near mileage stake number 5km).

[0098] Increase the redundancy of attitude sensors (such as dual IMU configuration) to improve the reliability of attitude data.

[0099] Acquisition strategy optimization: Increase the number of repeated acquisitions (such as flying back and forth 2 times) in the low-quality point section (such as mileage stake number 5km) to ensure data coverage.

[0100] Adjust sensor parameters (such as camera exposure time, lidar point cloud density) to improve data quality.

[0101] Both deviations and quality issues are related to the pipeline position. Causal relationships can only be established through analysis in the same spatial dimension. Avoid looking at deviations and quality issues in isolation to improve the pertinence of optimization. Through correlation analysis, the causal chain between deviations and quality issues can be identified to avoid blind optimization. Collaboratively optimize from three dimensions: path, attitude, and acquisition to ensure flight stability and data quality.

[0102] Step S104: Output the detection result of the pipeline by integrating the first detection information and the second detection information of the pipeline.

[0103] In some embodiments of the present application, outputting the detection result of the pipeline by integrating the first detection information and the second detection information of the pipeline includes: Determine the intersection information part and the difference set information part between the first detection information and the second detection information according to the position of the pipeline; Fuse the intersection information part of the first detection information and the second detection information to obtain fused information; Output the detection result of the pipeline based on the difference set information part and the fused information.

[0104] In this embodiment, based on the actual position of the pipeline (such as mileage stake number, GPS coordinates), map the first detection information and the second detection information to the same spatial coordinate system.

[0105] Definition of the intersection information part: The pipeline section covered by both detections (such as mileage stake number 0km - 10km).

[0106] Definition of the difference set information part: The pipeline section covered only by the first or the second detection (such as mileage stake number 10km - 15km covered only by the first detection).

[0107] Data weighted fusion: Assign weights to the detection results of the same section (for example, the weight of the second detection is higher because the quality is more reliable after optimization).

[0108] Processing of the difference set information: For the section covered only by the first detection (such as mileage stake number 10km - 15km): directly adopt the first detection result.

[0109] Process of outputting the detection result of the pipeline according to the detection information: The pipeline information collected by drones usually includes image data (visible light / infrared / lidar), position data (GPS / RTK), attitude data (roll / pitch / yaw angle), etc. The analysis process needs to combine multi-source data and obtain the detection conclusion through four steps: preprocessing, feature extraction, defect identification, and result integration.

[0110] 1. Data preprocessing Image data: Denoising: Gaussian filtering and median filtering are used to remove image noise.

[0111] Geometric correction: Correct image distortion according to the drone attitude data (such as pitch angle deviation).

[0112] Radiometric correction: Compensate for uneven illumination (such as solar radiation interference in infrared images).

[0113] Position data: Coordinate transformation: Convert GPS coordinates to pipeline mileage stakes for easy alignment with historical data.

[0114] Trajectory smoothing: Correct abnormal points in the flight trajectory (such as jumps caused by GPS signal loss) through Kalman filtering.

[0115] 2. Feature extraction and defect identification Image feature extraction: Texture feature: Use the gray-level co-occurrence matrix (GLCM) to extract the pipeline surface texture (such as the roughness of the corroded area).

[0116] Geometric feature: Identify geometric defects such as cracks and deformations through edge detection (such as the Canny algorithm).

[0117] Deep learning model: Object detection: Use models such as YOLO and Faster R-CNN to identify pipeline defects (such as corrosion, cracks, and deformations).

[0118] Semantic segmentation: Segment the defect area through models such as U-Net and DeepLab and calculate the area ratio.

[0119] Multi-source data fusion: Lidar point cloud: Extract pipeline three-dimensional deformation data (such as diameter change and curvature).

[0120] Infrared thermal imaging: Identify leakage points (such as temperature anomalies caused by natural gas leakage).

[0121] 3. Defect classification and quantification Defect classification: By type: corrosion, crack, deformation, leakage, etc.

[0122] By severity: Minor: The proportion of the defect area is less than 5%, and there is no structural risk.

[0123] Medium: 5% ≤ the proportion of the defect area < 15%, and regular monitoring is required.

[0124] Severe: The proportion of the defect area is ≥ 15%, and immediate repair is required.

[0125] Quantitative indicators: Corrosion depth: Calculated by 3D reconstruction of images or lidar point cloud computing.

[0126] Crack length: Calculated by image edge detection and geometric measurement.

[0127] 4. Result integration and report generation Spatial alignment: Align the defect location with the pipeline mileage stake number to generate a "defect - location" mapping table.

[0128] Correspondingly, the present application also provides an unmanned aerial vehicle - based pipeline detection system, as Figure 2 shown, including, The first module is used to obtain the pipeline information, pipeline location environment information, pipeline detection task, and unmanned aerial vehicle - carried sensor information of the pipeline to be measured, set the acquisition section of the pipeline according to the pipeline information and pipeline detection task of the pipeline to be measured, and formulate the first flight plan based on the pipeline location environment information and the acquisition section of the pipeline; The second module is used for the unmanned aerial vehicle to fly according to the first flight plan, collect the first detection information of the pipeline, count the flight information deviation between the first actual flight and the first flight plan, evaluate the quality of the first detection information, and determine whether a second flight is required based on the quality of the first detection information; The third module is used to optimize the second flight plan according to the flight information deviation if a second flight is required, and the unmanned aerial vehicle flies according to the second flight plan to collect the second detection information of the pipeline; The fourth module is used to output the detection result of the pipeline by integrating the first detection information and the second detection information of the pipeline.

[0129] Compared with the prior art, the beneficial effects of the present invention are: 1. Set the acquisition section of the pipeline according to the pipeline information and pipeline detection task of the pipeline to be measured, select reasonable unmanned aerial vehicle acquisition points considering the own situation of the pipeline to be measured, and provide a reliable basis for the subsequent flight plan planning of the unmanned aerial vehicle and pipeline detection. Formulate the first flight plan based on the pipeline location environment information and the acquisition section of the pipeline, and formulate the initial flight plan considering the pipeline location environment situation and the flight stability requirements, ensuring the adaptability of the unmanned aerial vehicle flight.

[0130] 2. Determine whether a second flight is required based on the quality of the first detection information. Judge whether the first flight is sufficient to support the pipeline detection quality requirements according to the quality of the detection information. When it is difficult to support, conduct a second flight. Optimize the second flight plan according to the flight information deviation, fill in the loopholes in the first flight, and ensure the reliability and adaptability of the second flight. Output the detection result of the pipeline by integrating the first detection information and the second detection information of the pipeline. Consider the complementarity and intersection between the first detection information and the second detection information to output the detection result of the pipeline, improve the adaptability and accuracy of the pipeline detection, ensure the adaptability of the UAV flight situation to the pipeline environment, and meet the detection requirements of the complex environment where the pipeline is located.

[0131] Through the description of the above embodiments, those skilled in the art can clearly understand that the present invention can be implemented by hardware or by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present invention can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.), including several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in various implementation scenarios of the present invention.

[0132] Those skilled in the art can understand that the drawings are only schematic diagrams of a preferred implementation scenario, and the modules or processes in the drawings are not necessarily essential for implementing the present invention.

[0133] Those skilled in the art can understand that the modules in the system in the implementation scenario can be distributed in the system of the implementation scenario according to the description of the implementation scenario, or can be correspondingly changed and located in one or more systems different from the present implementation scenario. The modules in the above implementation scenario can be combined into one module, or can be further split into multiple sub-modules.

[0134] The above is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution and inventive concept of the present invention, makes equivalent substitutions or changes, and all should be covered by the protection scope of the present invention.

Claims

1. A pipeline detection method based on an unmanned aerial vehicle, characterized in that including, obtaining the pipeline information, pipeline location environment information, pipeline detection task, and the information of sensors carried by the drone for the pipeline to be tested, setting the acquisition section of the pipeline according to the pipeline information and pipeline detection task of the pipeline to be tested, and formulating the first flight plan based on the pipeline location environment information and the acquisition section of the pipeline; the drone conducts flight according to the first flight plan, acquires the first detection information of the pipeline, counts the flight information deviation between the first actual flight and the first flight plan, evaluates the quality of the first detection information, and determines whether a second flight is required based on the quality of the first detection information; if a second flight is required, optimize the second flight plan according to the flight information deviation, and the drone conducts flight according to the second flight plan to acquire the second detection information of the pipeline; output the detection result of the pipeline by integrating the first detection information and the second detection information of the pipeline.

2. The method for detecting pipelines based on drones according to claim 1, wherein Setting the acquisition section of the pipeline according to the pipeline information and pipeline detection task of the pipeline to be tested, including, The pipeline information includes the pipeline length, pipeline orientation, pipeline material, and the medium transported inside the pipeline; Define the simple pipeline scenario and complex pipeline scenario according to the pipeline length and pipeline orientation. For the simple pipeline scenario, take the entire pipeline to be tested as the acquisition section of the pipeline to be tested; For the complex pipeline scenario, establish the pipeline model of the pipeline to be tested according to the pipeline length, pipeline orientation, pipeline material, and the medium transported inside the pipeline, collect all risk elements through the pipeline detection task, mark the high-risk positions at the corresponding positions of the pipeline model of the pipeline to be tested by integrating all risk elements, and take the high-risk positions as the acquisition section of the pipeline to be tested.

3. The method for detecting pipelines based on drones according to claim 2, wherein, Define the simple pipeline scenario and complex pipeline scenario according to the pipeline length and pipeline orientation, including, Extract multiple key features on the pipeline orientation, divide the pipeline into multiple sections according to the pipeline length, count the key feature density under each section to obtain the average key feature density, define the scenario complexity based on the pipeline length, multiple key features, and the average key feature density, and distinguish the simple pipeline scenario and complex pipeline scenario through the scenario complexity.

4. The method for detecting pipelines based on drones according to claim 1, wherein Formulate the first flight plan based on the pipeline location environment information and the acquisition section of the pipeline, including, Import the pipeline information and pipeline location environment information into the GIS software, and overlay the obstacle distribution map in the pipeline location environment. Mark the positions of the acquisition section of the pipeline in the GIS software. Take the positions of multiple acquisition sections as waypoints, and generate multiple candidate paths through the path planning algorithm. Each candidate path passes through all waypoints; Split each candidate path into multiple path units, conduct path cost evaluation and flight risk cost evaluation for each candidate path based on the path units, generate the comprehensive cost by combining the path cost and the flight risk cost, and draw the comprehensive cost change curve that changes with the path units; Distinguish the stable part curve and the fluctuating part curve on the comprehensive cost change curve, calculate the relative cost of the stable part curve and the fluctuating part curve respectively through the change of the curve slope, so as to determine the relative comprehensive cost, and select a candidate path as the target path by virtue of the relative comprehensive cost; Configure the flight information of the drone based on the target path.

5. The method for detecting pipelines based on an unmanned aerial vehicle according to claim 4, wherein Configure the flight information of the drone based on the target path, including, divide the target path into a flight path and a collection path according to the collection section position of the pipeline; For the flight path, analyze the environmental requirements for flight stability through GIS software, and configure the flight information of the drone in combination with the flight mission requirements; For the collection path, analyze the pipeline's requirements for flight stability and the environment's requirements for flight stability through GIS software, and configure the flight information of the drone based on the pipeline's requirements for flight stability, the environment's requirements for flight stability, and the flight mission requirements.

6. The method for detecting pipelines based on drones according to claim 5, characterized in that, Statistically analyze the flight information deviation between the first actual flight and the first flight plan, including, respectively statistically analyze the deviation of the flight information for the flight path and the collection path to obtain the first deviation and the second deviation, and plot the respective change curves of the first deviation and the second deviation with respect to position. Identify the respective deviation points on the change curves of the first deviation and the second deviation, and analyze the correlation between the two types of deviation points.

7. The method for pipeline detection based on an unmanned aerial vehicle according to claim 6, wherein Optimize the second flight plan according to the flight information deviation, including, spatially align the flight information deviation and the quality of the first detection information through the pipeline position, and optimize to obtain the second flight plan based on the deviation points and the correlation between the deviation points.

8. The method for pipeline detection based on an unmanned aerial vehicle according to claim 4, characterized in that, Output the detection result of the pipeline by integrating the first detection information and the second detection information of the pipeline, including, determine the intersection information part and the difference set information part between the first detection information and the second detection information according to the position of the pipeline; fuse the intersection information part of the first detection information and the second detection information to obtain the fused information; output the detection result of the pipeline based on the difference set information part and the fused information.

9. An unmanned aerial vehicle-based pipeline detection system, characterized in that, including, The first module is used to obtain the pipeline information, pipeline position environment information, pipeline detection task, and drone-borne sensor information of the pipeline to be measured. Set the collection section of the pipeline according to the pipeline information and pipeline detection task of the pipeline to be measured, and formulate the first flight plan based on the pipeline position environment information and the collection section of the pipeline; The second module is used for the drone to fly according to the first flight plan, collect the first detection information of the pipeline, statistically analyze the flight information deviation between the first actual flight and the first flight plan, evaluate the quality of the first detection information, and judge whether a second flight is required based on the quality of the first detection information; The third module is used, if a second flight is required, to optimize the second flight plan according to the flight information deviation, and the drone flies according to the second flight plan to collect the second detection information of the pipeline; The fourth module is used to output the detection result of the pipeline by integrating the first detection information and the second detection information of the pipeline.

Images