An intelligent identification system for abnormal hidden dangers in pipeline inspection based on drones

Abstract

The present invention relates to the field of intelligent inspection technology, and specifically to an intelligent identification system for abnormal hidden dangers in pipeline inspections based on drones. In the present invention, by constructing a three-dimensional grid for the data collected by the drone during the inspection process, the spatial division of the inspection area is achieved, so that the detection of gas pipelines has higher accuracy. Combined with hyperspectral data analysis, the determination of methane concentration can be accurately mapped to spatial grid points, reducing the interference of environmental factors and improving the accuracy of preliminary identification of leakage areas. Based on the calculation of local concentration gradients, through data smoothing and diffusion optimization methods, the boundaries of the leakage area are made clearer, the interference of environmental noise is eliminated, and the concentration value is dynamically adjusted to enhance the adaptability to the changing trend of the leakage point. With the construction of a virtual diffusion field, combined with the spatial occupancy of terrain obstacles, the analysis of the gas flow path is made more accurate, solving the problem of misjudgment caused by the complexity of the terrain.

Description

Technical Field

[0001] The present invention relates to the field of intelligent inspection technology, and in particular to an intelligent identification system for abnormal hidden dangers in pipeline inspection based on unmanned aerial vehicles. Background Art

[0002] The drone-based pipeline inspection and intelligent identification system for abnormalities and hidden dangers uses drones equipped with high-precision sensing equipment to inspect gas pipelines and intelligently identify abnormalities. This system addresses issues such as gas leaks, pipeline damage, and the impact of geological disasters during pipeline inspections. It uses drones equipped with laser methane telemetry equipment, high-definition optical imaging equipment, and infrared thermal imagers to obtain high-resolution data from the inspection area. Combined with computer vision algorithms, the system analyzes the spectral characteristics of gas leak areas, performs morphological matching on structural changes on the pipeline surface, and performs differential calculations on the ambient temperature gradient to identify pipeline abnormalities and hidden dangers.

[0003] Existing inspection methods are limited by the coverage of sensors during data collection, and inspection accuracy relies on data collection at a single point or a small number of points, resulting in data breakpoints during the inspection process, making it difficult to accurately capture leaks in some areas. Data processing relies on the judgment method of a single concentration value, which is easily affected by environmental factors, resulting in a high misjudgment rate. Especially in environments with large wind speed changes, the gas diffusion pattern is complex, and it is difficult to form an accurate judgment based solely on the measured concentration value. The lack of consideration of terrain obstacles during the inspection process leads to restricted gas flow in complex terrain areas but cannot be correctly analyzed, increasing the possibility of missed detections. The prediction of gas diffusion trends mostly relies on experience-based judgments and lacks intelligent analysis based on historical data, resulting in insufficient early warning capabilities and difficulty in predicting possible hidden dangers in advance, causing delayed inspection responses and affecting the accuracy and timeliness of pipeline maintenance work. Summary of the Invention

[0004] The purpose of the present invention is to solve the shortcomings of the existing technology and propose an intelligent identification system for pipeline inspection abnormalities and hidden dangers based on drones.

[0005] In order to achieve the above objectives, the present invention adopts the following technical solutions: A pipeline inspection abnormality hidden danger intelligent identification system based on drones includes:

[0006] The methane concentration detection module collects the UAV flight data to construct a first grid in the three-dimensional space, evenly divides the first grid into a plurality of first grid points, and maps the methane concentration value of each first grid point according to the measured laser absorption spectrum data to generate a methane concentration heat map;

[0007] The concentration data smoothing module reads all methane concentration values ​​in the methane concentration heat map and calculates the local methane concentration gradient at each first grid point. Based on the magnitude of the local methane concentration gradient, it marks the potential leakage area and sets the diffusion coefficient corresponding to the first grid point to update the methane concentration value, thus generating a diffusion-optimized methane concentration heat map.

[0008] The virtual diffusion field reconstruction module constructs a second grid point occupied by the terrain obstacle space on the diffusion-optimized methane concentration thermodynamic map, sets a methane concentration attenuation coefficient for the second grid point occupied by the obstacle space, analyzes the methane concentration value from the first grid point corresponding to the potential leakage area to the second grid point occupied by the terrain obstacle space based on the methane concentration attenuation coefficient, and generates a methane concentration distribution map after terrain analysis;

[0009] The leakage trend prediction module predicts the change trend of methane around the potential leakage area in the methane concentration distribution map after the terrain analysis, and generates a diffusion trend prediction result;

[0010] The leakage risk assessment module identifies the change trend of methane in the diffusion trend prediction result and visually displays it on the methane concentration distribution map after the terrain analysis, thereby generating a gas pipeline leakage hazard identification result.

[0011] As a further solution of the present invention, the steps for obtaining the methane concentration thermodynamic map are specifically as follows:

[0012] Collecting drone flight data, including drone location information and flight trajectory data along the gas pipeline area, determining the three-dimensional coordinates of all measurement points, integrating the three-dimensional coordinates of all measurement points to construct a first grid in the three-dimensional space, evenly dividing the first grid into a plurality of first grid points, and generating a first grid division result;

[0013] Acquiring terrain height data, adjusting the first grid points in the first grid division result to match the current terrain structure, collecting laser absorption spectrum data measured by a methane laser telemeter on a UAV, analyzing the methane concentration in the air based on the laser absorption spectrum data, and generating a methane concentration value;

[0014] The gaps in the methane concentration values ​​between the measurement points are filled by an interpolation method and mapped to each first grid point in the three-dimensional space to generate a methane concentration heat map.

[0015] As a further solution of the present invention, the steps for obtaining the methane concentration thermodynamic map after the diffusion optimization are specifically as follows:

[0016] Reading all methane concentration values ​​in the methane concentration thermodynamic map, calculating the local methane concentration gradient at each first grid point, comparing the local methane concentration gradient with a preset change threshold, marking potential leakage areas where the local methane concentration gradient exceeds the change threshold, and generating a potential leakage area marking result;

[0017] Based on the local methane concentration gradient of the first grid point corresponding to each potential leakage area in the potential leakage area marking result, the diffusion coefficient of the first grid point corresponding to the potential leakage area is set, and the methane concentration value of the first grid point corresponding to the potential leakage area is updated through the partial differential diffusion equation according to the diffusion coefficient to generate a diffusion-optimized methane concentration thermodynamic map.

[0018] As a further solution of the present invention, the steps for obtaining the methane concentration distribution map after the terrain analysis are specifically as follows:

[0019] Collecting three-dimensional terrain point cloud data scanned by UAV LiDAR, constructing a second grid point occupied by the terrain obstacle space on the diffusion-optimized methane concentration heat map based on the three-dimensional terrain point cloud data, and discretizing the methane diffusion path around the potential leakage area using the finite volume method with reference to the position and distribution of the second grid point to generate a methane diffusion path processing result;

[0020] Obtain real-time wind speed, wind direction, temperature, and humidity data for the terrain obstacle area corresponding to the methane diffusion path processing results, analyze the total flow deviation of methane in the terrain obstacle area, and set the methane concentration attenuation coefficient of the second grid point occupied by the obstacle space based on the total flow deviation;

[0021] The methane concentration values ​​from the first grid point corresponding to the potential leakage area to the second grid point occupied by the terrain obstacle space are analyzed according to the methane concentration attenuation coefficient to generate a methane concentration distribution map after terrain analysis.

[0022] As a further solution of the present invention, the steps for obtaining the diffusion trend prediction result are specifically as follows:

[0023] Identifying the diffusion direction of methane around the potential leakage area in the methane concentration distribution map after the terrain analysis, and generating a methane diffusion direction analysis result;

[0024] Based on the methane diffusion direction analysis results, all records of historical methane leakage are obtained, and the changing trend of methane concentration in the future is predicted in combination with turbulent diffusion theory to generate a diffusion trend prediction result.

[0025] As a further solution of the present invention, the steps for obtaining the gas pipeline leakage hidden danger identification result are specifically as follows:

[0026] Visually display the methane concentration change trend in the future time in the diffusion trend prediction result on the methane concentration distribution map after the terrain analysis, identify the accumulation of methane concentration in different potential leakage areas, and obtain a methane concentration evolution display map;

[0027] Based on the methane concentration change trend in the methane concentration evolution display diagram, the risk level of the current potential leakage area is evaluated, and a gas pipeline leakage hazard identification result is generated.

[0028] Compared with the prior art, the advantages and positive effects of the present invention are:

[0029] In this invention, a three-dimensional grid is constructed based on data collected by drones during inspections, achieving spatial division of the inspection area and achieving higher accuracy in gas pipeline inspections. Combined with hyperspectral data analysis, methane concentration measurements can be precisely mapped to spatial grid points, reducing interference from environmental factors and improving the accuracy of initial identification of leak areas. Based on local concentration gradient calculations, data smoothing and diffusion optimization methods are used to clarify the boundaries of leak areas, eliminating interference from environmental noise. Concentration values ​​are dynamically adjusted to enhance adaptability to changing leak trends. By constructing a virtual diffusion field and taking into account the spatial occupancy of terrain obstacles, gas flow path analysis is more accurate, addressing the problem of misjudgment caused by terrain complexity. Incorporating real-time analysis of environmental parameters, the diffusion path calculation method is dynamically adjusted based on the influence of wind speed, direction, temperature, and humidity, ensuring that the predicted methane concentration distribution trend is more accurate. Incorporating turbulent diffusion theory, correlation analysis of historical leak data is performed to optimize the accuracy of future leak trend predictions, providing more reliable data support for inspection decisions. The concentration evolution in the leakage area is visualized, allowing inspection personnel to intuitively judge the severity of potential hidden dangers. Combined with risk assessment methods, dynamic classification of leakage hazards can be achieved, improving the pertinence of early warnings and response speed. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 is a system flow chart of the present invention;

[0031] Figure 2 A flow chart for obtaining a methane concentration thermodynamic map according to the present invention;

[0032] Figure 3 Flowchart for obtaining a methane concentration thermodynamic map after diffusion optimization according to the present invention;

[0033] Figure 4 A flow chart of the present invention for obtaining a methane concentration distribution map after terrain analysis;

[0034] Figure 5 A flow chart for obtaining diffusion trend prediction results according to the present invention;

[0035] Figure 6 This is a flow chart of the present invention for obtaining the results of identifying hidden dangers of gas pipeline leakage. DETAILED DESCRIPTION

[0036] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0037] See also Figure 1 , a pipeline inspection abnormal hidden danger intelligent identification system based on drones includes:

[0038] The methane concentration detection module collects the UAV flight data to construct a first grid in the three-dimensional space, evenly divides the first grid into a plurality of first grid points, and maps the methane concentration value of each first grid point according to the measured laser absorption spectrum data to generate a methane concentration heat map;

[0039] The concentration data smoothing module reads all methane concentration values ​​in the methane concentration heat map and calculates the local methane concentration gradient at each first grid point. Based on the magnitude of the local methane concentration gradient, it marks the potential leakage area and sets the diffusion coefficient corresponding to the first grid point to update the methane concentration value, thus generating a diffusion-optimized methane concentration heat map.

[0040] The virtual diffusion field reconstruction module constructs a second grid point occupied by the terrain obstacle space on the diffusion-optimized methane concentration heat map, sets the methane concentration attenuation coefficient for the second grid point occupied by the obstacle space, and analyzes the methane concentration values ​​from the first grid point corresponding to the potential leakage area to the second grid point occupied by the terrain obstacle space based on the methane concentration attenuation coefficient to generate a methane concentration distribution map after terrain analysis.

[0041] The leakage trend prediction module predicts the changing trend of methane around the potential leakage area in the methane concentration distribution map after terrain analysis and generates the diffusion trend prediction result;

[0042] The leakage risk assessment module identifies the changing trend of methane in the diffusion trend prediction results and visualizes it on the methane concentration distribution map after terrain analysis, generating the gas pipeline leakage hazard identification results.

[0043] See also Figure 2 , the steps for obtaining the methane concentration heat map are as follows:

[0044] Collecting drone flight data, including drone location information and flight trajectory data along the gas pipeline area, determining the three-dimensional coordinates of all measurement points, integrating the three-dimensional coordinates of all measurement points to construct a first grid in the three-dimensional space, evenly dividing the first grid into a plurality of first grid points, and generating a first grid division result;

[0045] The high-precision GPS module (such as Trimble RTK-GNSS) and inertial measurement unit (IMU) carried by the drone are used to obtain the location information of the drone when it flies in the gas pipeline area, and the heading angle, acceleration and angular velocity data in the drone flight control system log are recorded. These data are stored in real time by the drone control system and can be extracted by flight control software (such as Pix4D or DJI Terra). During the flight of the drone, the geographic coordinate information of each measurement point collected includes longitude, latitude and altitude. These coordinate data are parsed by geographic information processing software (such as ArcGIS or QGIS), and the coordinate conversion tool is used to convert the geographic coordinates into X, Y, and Y coordinates in the rectangular coordinate system. During the conversion process of Z three-dimensional coordinates, the World Geographic Coordinate System (WGS84) standard is referenced to ensure consistent coordinate accuracy. Subsequently, the three-dimensional coordinate data of all measurement points are imported into point cloud processing software (such as CloudCompare or MeshLab). The spatial distribution of the measurement points is determined through point cloud density analysis. On this basis, the Delaunay triangulation method is used to connect the measurement points and construct the first grid. When dividing the first grid, the division parameters are set in the point cloud processing software according to the size of the measurement area and the density of the measurement points. The software automatically generates uniform grid units and exports the divided grid data. After the division is completed, the first grid serves as the basic data structure of the three-dimensional space.

[0046] Obtain terrain height data, adjust the first grid point in the first grid division result to match the current terrain structure, collect laser absorption spectrum data measured by the methane laser telemeter on the UAV, analyze the methane concentration in the air based on the laser absorption spectrum data, and generate a methane concentration value;

[0047] Combined with the terrain height data obtained by the drone, first use terrain data processing software (such as GlobalMapper or Surfer) to load and compare the digital elevation model (DEM) data of the measurement area, analyze the height difference between the measurement point and the DEM data, and use GIS software (such as ArcGIS's terrain analysis tool) to overlay the DEM data on the first grid, and use the spatial adjustment tool to match the height of the first grid to ensure that the measurement point data conforms to the actual terrain conditions. During the adjustment process, the bilinear interpolation method is used to make a smooth transition between the height of each grid cell in the first grid and the surrounding known terrain data. After the adjustment is completed, the drone is equipped with a methane stimulator. An optical remote sensing instrument (such as Pergam Laser Methane mini or SPM LiDAR) flies along the gas pipeline area and uses laser to scan the methane concentration in the air. The device performs remote detection through a laser beam of a specific wavelength. The detection data is read by methane detection software (such as Pergam data analysis software or Laser Methane Viewer) and automatically generates a methane concentration dataset for the measurement point. These concentration data are matched with the coordinate data after the first grid adjustment through data matching software (such as MATLAB or Python's GDAL library) and integrated into a methane concentration data file in three-dimensional space.

[0048] The gaps in the methane concentration values ​​between the measurement points are filled by interpolation and mapped to each first grid point in the three-dimensional space to generate a methane concentration heat map;

[0049] After obtaining the methane concentration data at the measurement points, first use the interpolation tools provided by GIS software (such as ArcGIS or Surfer) to interpolate and fill in the gaps between data points. The interpolation method can be inverse distance weighted interpolation (IDW) or kriging interpolation (Kriging). In ArcGIS, use the interpolation analysis tool to input the methane concentration values ​​of known measurement points. The software automatically calculates and fills the concentration data of unknown measurement points. After interpolation is complete, the generated complete concentration data is imported into 3D visualization software (such as ParaView or Tecplot360). Methane concentration mapping is performed on the first grid, and the concentration values ​​correspond one-to-one with the grid cells. A 3D heat map is generated. The heat map intuitively presents the spatial distribution of methane concentration through color changes. Finally, it is exported to a standard GIS format file or visualization format.

[0050] Using drones to collect flight data, a detailed three-dimensional grid is constructed, achieving high-precision spatial division of the inspection area, thereby improving the accuracy of gas pipeline inspections. Terrain height data is incorporated into the grid structure to adjust the measurement results, ensuring that they better reflect the actual terrain characteristics and enhancing the authenticity and reliability of the data. A methane laser telemeter provides hyperspectral data, and interpolation is used to fill in missing values ​​to ensure the integrity of the concentration distribution. Finally, methane concentration values ​​are mapped using the three-dimensional grid and a heat map is generated to accurately reflect the spatial distribution of gas leaks.

[0051] See also Figure 3 , the steps for obtaining the methane concentration heat map after diffusion optimization are as follows:

[0052] Read all methane concentration values ​​in the methane concentration thermodynamic map, calculate the local methane concentration gradient at each first grid point, compare the local methane concentration gradient with a preset change threshold, mark potential leakage areas where the local methane concentration gradient exceeds the change threshold, and generate a potential leakage area marking result;

[0053] For calculating the local methane concentration gradient at each first grid point, the formula is used:

[0054]

[0055] The local methane concentration gradient G( i,j );

[0056] Where C is the methane concentration value of the first grid point in the methane concentration thermodynamic map, in ppm, which is collected by the methane sensor; is the partial derivative of methane concentration along the x-direction, indicating the concentration trend within a unit distance. This value is calculated using the finite difference method: is the partial derivative of methane concentration along the y direction, which indicates the concentration trend within a unit distance. This value is calculated by the finite difference method: (i, j) are the coordinates of the current first grid point, indicating the number of the current grid cell. This value is automatically set by the grid division rule and is generally obtained based on the geographic location or sensor layout plan. i+1 and i-1 represent the adjacent grid points to the right and left of the current first grid point coordinate (i, j), respectively. j+1 and j-1 represent the adjacent grid points above and below the current first grid point coordinate (i, j), respectively. Δx and Δy are the grid steps (in meters) along the x and y directions of the current first grid point, determined by the grid division plan, for example, a grid spacing of 1 meter. They are usually obtained from a GIS (Geographic Information System) or sensor layout plan.

[0057] Assume that the methane concentration C(i,j) and the adjacent point concentration C(i,j) = 5 ppm, C(i+1,j) = 12 ppm (methane concentration to the right of the point), C(i-1,j) = 3 ppm (methane concentration to the left of the point), C(i,j+1) = 11 ppm (methane concentration above the point), and C(i,j-1) = 2 ppm (methane concentration below the point).

[0058] Calculation process:

[0059]

[0060] The results show that the current methane concentration change at the first grid point is 6.36 ppm / m.

[0061] Set the change threshold G th To identify possible methane leakage areas, the threshold is set based on laboratory gas diffusion tests, historical leakage data, environmental parameters (temperature, humidity, wind speed) and sensitivity analysis of monitoring equipment. The value range is usually between 3.0ppm / m3 and 4.0ppm / m3. In this calculation, G is selected. th =3.5ppm / m3 as the initial setting value, the purpose is to distinguish the normal diffusion area from the abnormal concentration change area. Under normal environmental conditions, the change of methane concentration gradient is usually relatively stable. Lower concentration gradients (such as less than 2ppm / m3) often appear in the stable diffusion area, while higher concentration gradients (such as greater than 3.5ppm / m3) usually mean that the local concentration has changed dramatically, which may be caused by point source leakage, local concentration accumulation, or external disturbances (such as wind speed changes). Therefore, when the methane concentration gradient G( i,j ) exceeds 3.5ppm / m, it is considered that there may be potential leakage in the area and further analysis of its diffusion is required. When the concentration gradient is lower than the threshold, the area is considered to be a stable concentration area and no immediate leakage treatment is required. Combined with the calculation results, the concentration gradient G( i,j )=6.36ppm / m, far exceeding the set threshold G th =3.5ppm / m, so this grid cell is marked as a potential leakage area, which serves as the basis for setting the subsequent diffusion coefficient to further calculate the diffusion trend of methane concentration in this area.

[0062] Based on the local methane concentration gradient of the first grid point corresponding to each potential leakage area in the potential leakage area marking result, the diffusion coefficient of the first grid point corresponding to the potential leakage area is set. The methane concentration value of the first grid point corresponding to the potential leakage area is updated according to the diffusion coefficient using the partial differential diffusion equation to generate a diffusion-optimized methane concentration heat map.

[0063] The diffusion coefficient D(i,j) of methane reflects the diffusion ability of gas in space. The value of the diffusion coefficient is usually determined by experiment or set by referring to meteorological parameters and literature data. Under normal atmospheric conditions, the molecular diffusion coefficient of methane in air is about 0.1 to 0.3 m 2 / s, which may increase to 0.5-1.0m under the influence of strong wind or turbulence 2 / s. In this calculation, the diffusion coefficient is set based on the local methane concentration gradient G(i,j) at the grid point. A higher concentration gradient usually means a more drastic concentration change, and the corresponding diffusion capacity may be enhanced. Therefore, the diffusion coefficient is set in sections to make the diffusion calculation more in line with the actual situation. The setting rules are as follows: when G(i,j)<2ppm / m, it is considered that the concentration change in this area is gentle, and the diffusion coefficient takes the smaller value D(i,j)=0.1m 2 / s; when 2≤G(i,j)<3.5ppm / m, the concentration change is considered moderate, and the diffusion coefficient is D(i,j)=0.2m 2 / s; When G(i,j)≥3.5ppm / m3, it is considered that the concentration changes dramatically, and there may be a leakage source or a high diffusion area. The diffusion coefficient takes the larger value D(i,j)=0.6m 2 / s, combined with the calculation results, the concentration gradient of the current grid point G(i,j) = 6.36ppm / m, which exceeds the set high gradient threshold G th =

[0064] 3.5ppm / m, so the diffusion coefficient of this grid point is set to D(i,j)=0.6m 2 / s to simulate the temporal variation trend of methane concentration in the area.

[0065] To update the methane concentration value of the first grid point corresponding to the potential leakage area through the partial differential diffusion equation, the formula is used:

[0066]

[0067] Calculate the updated methane concentration value C at the first grid point (i, j) corresponding to the potential leakage area new (i,j);

[0068] Where C(i,j) is the methane concentration value corresponding to the first grid point (i,j) in the potential leakage area, which is obtained by sensor monitoring; D(i,j) is the diffusion coefficient set according to the first grid point corresponding to the potential leakage area (unit: m 2 / s), which is obtained based on experimental measurements and gas diffusion characteristics; Δx, Δy are the grid steps along the x and y directions corresponding to the first grid point in the potential leakage area (unit: m). In a uniform grid, Δx = Δy, but it needs to be calculated separately in a non-uniform grid; Δt is the time step set according to the grid step, which is used to discretize the time derivative and is usually selected according to the simulation accuracy requirements, such as 1s, 0.5s, etc.; S x is the Laplace term of the methane concentration along the x-axis corresponding to the first grid point in the potential leakage area, which is used to calculate the diffusion effect along the x-axis. x =C(i+1,j)+C(i-1,j)-2C(i,j);S y is the Laplace term of the methane concentration along the y-axis corresponding to the first grid point in the potential leakage area, which is used to calculate the diffusion effect along the y-axis. y =C(i,j+1)+C(i,j-1)-2C(i,j), (i,j) is the coordinate of the first grid point corresponding to the potential leakage area, which represents the number of the grid cell corresponding to the potential leakage area. This value is automatically set by the grid division rule and is generally obtained based on the geographical location or sensor layout plan; i+1 and i-1 represent the adjacent grid points to the right and left of the first grid point (i,j) corresponding to the potential leakage area; j+1 and j-1 represent the adjacent grid points above and below the first grid point (i,j) corresponding to the potential leakage area, respectively. 2C(i,j) represents twice the methane concentration at the first grid point (i,j) corresponding to the potential leakage area. Its function is to ensure that the current point concentration forms a dynamic equilibrium with the concentration of its adjacent points in the diffusion calculation, and to determine the concentration change trend of this point in the diffusion calculation; and Both factors are dimensionless and represent the proportion of the contribution of diffusion to the concentration update during that time step.

[0069] Assume that C(i,j) = 5ppm, C(i+1,j) = 12ppm (right grid), C(i-1,j) = 3ppm (left grid), C(i,j+1) = 11ppm (upper grid), C(i,j-1) = 2ppm (lower grid), D(i,j0 = 0.6m 2 / s, Δt=1s, Δx=Δy=1m.

[0070] Calculate S x :S x =C(i+1,j)+C(i-1,j)-2C(i,j)=12+3-2(5)=5;

[0071] Calculate S y :S y =C(i,j+1)+C(i,j-1)-2C(i,j)=11+2-2(5)=3;

[0072] Calculate C new (i,j):

[0073] The results show that the methane concentration at this grid point increased from 5 ppm to 9.8 ppm within a single time step, indicating that diffusion caused the concentration to converge at this point, which is consistent with the characteristics of a high-gradient diffusion region. Subsequent iterative calculations can be performed over multiple time steps to further analyze the changing trend of methane concentration and determine the potential location of the leak source. The updated methane concentration data is remapped to the original 3D grid, ensuring that the new concentration value corresponds to each grid cell. The updated concentration data is then imported into visualization software in CSV, VTK, or NetCDF format to generate a diffusion-optimized methane concentration heat map.

[0074] By calculating the methane concentration gradient, we can precisely identify potential leak areas, effectively avoiding misjudgments caused by environmental interference and improving leak detection accuracy. By setting the diffusion coefficient and optimizing concentration data using the partial differential diffusion equation, we can make the gas diffusion pattern more realistic and enhance dynamic adaptability to leak trends. Furthermore, the optimized heat map clearly displays the concentration distribution, improving data readability and decision-making efficiency, providing strong support for accurately locating leaks and optimizing inspection strategies.

[0075] See also Figure 4 The steps for obtaining the methane concentration distribution map after terrain analysis are as follows:

[0076] Collect three-dimensional terrain point cloud data from drone LiDAR scans. Based on this data, construct a second grid of points representing the space occupied by terrain obstacles on the diffusion-optimized methane concentration heat map. Based on the location and distribution of these second grid points, use the finite volume method to discretize the methane diffusion paths around the potential leakage area, generating a methane diffusion path processing result.

[0077] First, the drone's flight altitude, scanning frequency, and laser emission angle are set to ensure coverage of the terrain information of the potential leakage area. During the flight, the LiDAR device emits laser pulses and measures the time difference of the return signal to obtain the three-dimensional coordinate information of the terrain surface. After data collection, the point cloud is corrected by combining the drone's inertial measurement unit (IMU) and global navigation satellite system (GNSS) data to ensure that the spatial positioning error is within the set threshold range. Subsequently, a filtering method is used to remove abnormal points, and the terrain obstacle area is extracted based on the distribution characteristics of the point cloud. The judgment standard is the joint threshold setting of the point cloud density and elevation change. If the point cloud density in a certain area exceeds 500 points / m 2, and the elevation change exceeds 2m, it is determined that there is a terrain obstacle in the area. According to the spatial occupancy of the point cloud data, the second grid point occupied by the terrain obstacle space is constructed on the methane concentration heat map after diffusion optimization. The grid is divided into fixed 1m×1m×1m units. The occupancy state of each unit is determined by the point cloud density. If the point cloud density exceeds the set threshold, it is marked as an obstacle, otherwise it is marked as idle. After the grid construction is completed, the finite volume method (FVM) is used to discretize the methane diffusion path around the potential leakage area based on the position and distribution of the second grid point. First, the calculation area is divided and the boundary conditions between the grids are set. Inside each grid cell, the mass conservation equation is applied to calculate the gas concentration, and the flux exchange between the grids is adjusted according to the wind speed, wind direction and terrain influence. Finally, the discretized diffusion path data is stored.

[0078] Obtain real-time wind speed, wind direction, temperature, and humidity data for the terrain obstacle area corresponding to the methane diffusion path processing results, analyze the total flow deviation of methane in the terrain obstacle area, and set the methane concentration attenuation coefficient of the second grid point occupied by the obstacle space based on the total flow deviation;

[0079] To analyze the total flow deviation of methane in the terrain obstacle area, the formula is used: Calculate the normalized deviation ΔD of the wind speed in the current terrain obstacle area relative to the historical wind speed range in the area V , the value range is 0 to 1. The larger the value, the stronger the effect of wind speed on methane diffusion. If ΔD V If it is close to 1, it means that the current wind speed is close to the historical maximum wind speed in the area, which is conducive to methane diffusion. V Close to 0, indicating low wind speed and weak diffusion capacity; V is the real-time wind speed, obtained through wind speed sensor or weather station; V min , V max They are the minimum and maximum wind speeds around the terrain obstacle area (unit: m / s). The wind speed range in the area is calculated based on historical meteorological data. Data is usually obtained from meteorological databases (such as NOAA or local weather monitoring stations), and the data of the past year is analyzed to determine the extreme value range.

[0080] Using the formula: ΔD θ =1-cos(θ-θ0); calculate the deviation degree ΔD between the wind direction in the current terrain obstacle area and the main diffusion direction of methane θ , the value range is 0 to 2. The larger the value, the greater the wind direction deviation, and the more the dominant direction of methane diffusion is disturbed. θ =0, indicating that the wind direction is completely consistent, which is conducive to the spread of methane along the main diffusion path; if ΔD θ≈2, indicating that the wind direction is completely opposite and diffusion is strongly suppressed; where θ is the real-time wind direction (unit: °), obtained through a wind direction sensor or weather station, and θ0 is the main diffusion direction (unit: °). Based on the long-term average dominant wind direction of the methane leak point and the surrounding airflow, combined with historical wind direction data statistics, a wind rose diagram can be used to analyze wind direction changes in the past month or year, with the main wind direction taken as θ0.

[0081] Using the formula: Calculate the normalized deviation ΔD between the current terrain obstacle area temperature and the preset reference temperature T The value range is 0 to 1. The larger the value, the more obvious the effect of temperature on diffusion. The main way in which temperature affects diffusion is to change the air density and turbulence. If ΔD T =0, indicating that the temperature is close to the reference value and the temperature has little effect; if ΔD T =1, indicating that the temperature is in an extreme range, which may lead to abnormal diffusion patterns (such as thermal convection or inversion layer); where T is the real-time temperature (unit: K), obtained by temperature sensor or weather station, T ref It is the standard temperature (unit: K), which is set according to the environmental standard and is usually based on 298K (25℃). min , T max They are the minimum and maximum temperatures in the monitoring area (unit: K), which are based on the temperature range in the area calculated based on historical meteorological data, usually obtained from the meteorological database, and analyze the maximum and minimum temperatures in the past year.

[0082] Using the formula: Calculate the normalized deviation ΔD between the current terrain obstacle area humidity and the preset reference humidity H , the value range is 0 to 1. The larger the value, the greater the effect of humidity on diffusion. Humidity affects diffusion in several ways, including changing the buoyancy of the air and affecting the particle settling rate. If ΔD H =0, indicating that the humidity is close to the reference value and has little impact; if ΔD H =1, indicating that the humidity is in an extreme range, which may affect the detection of methane concentration; where H is the real-time humidity (unit: %), obtained through a humidity sensor or weather station, H ref It is the reference humidity (unit:%), usually 50% is used as the reference humidity, and the average humidity is used as the reference value. min , H max They are the minimum and maximum humidity in the monitoring area (unit: %), and the historical humidity range is analyzed from meteorological data, usually obtained from meteorological monitoring station data, to analyze the humidity changes in the past year.

[0083] Use the formula: ΔD'=w V ΔD V +w θΔD θ +w T ΔD T +w H ΔD H ; Calculate the total flow deviation ΔD'; where w V 、w θ 、w T 、w H Represent the relative influence weights of wind speed, wind direction, temperature and humidity on methane diffusion.

[0084] The methane concentration attenuation coefficient k is calculated using the formula: k = k0 × (1-ΔD); where k0 is the basic attenuation term preset according to the current terrain obstacle area (unit: m -1 ), determined by the characteristics of terrain obstacles.

[0085] Assume that the real-time monitoring data is: wind speed: V = 5.5m / s, wind direction: θ = 210°, temperature: T = 305K, humidity: H = 65%. Historical range data: V min =1.2, V max =8.0m / s,θ0=200°,T min =280, T max =320K, H min =30%, H max =90%. The basic attenuation coefficient is set as:

[0086] k0=0.08m -1 The basic attenuation coefficient k0 is mainly determined by the characteristics of terrain obstacles, reflecting the initial obstruction of obstacles to methane diffusion. Its setting is based on the following: if the obstacle density is high (such as high-rise buildings, mountains, etc.), the air circulation is limited, and the diffusion resistance is large, k0 takes a higher value (0.1-0.2m -1 When the density of obstacles is low (such as open plains, open waters, etc.), diffusion is less hindered and k0 takes a lower value (0.02-0.08m -1 Rough terrain (such as forests and hills) creates more turbulence, complicating the diffusion path and resulting in a higher k0 value. Flat terrain (such as deserts and airport runways) allows for smoother diffusion and a lower k0 value. Wind tunnel experiments were conducted to measure the decay of methane concentrations under different terrains and to fit the k0 value. Computational fluid dynamics (CFD) was used to simulate diffusion patterns under different wind speeds, temperatures, and humidity conditions, calculate k0 tailored to the specific environment, and calibrate it using actual measurement data.

[0087] Calculate wind speed deviation: Calculate wind direction deviation: ΔD θ =1-cos(210°-200°)=0.0152; calculate temperature deviation Calculating humidity deviation The calculation results show that wind speed has the greatest impact (deviation 0.632). Higher wind speed has a greater promoting effect on diffusion, so its weight is set to the highest (w V =0.3). The wind direction has the smallest influence (deviation 0.0152), which means that the wind direction is very close to the main diffusion direction, so its weight is the lowest (w θ =0.2). Temperature and humidity have moderate effects (deviations of 0.175 and 0.25), so they are both given higher weights (w T =0.25,w H =0.25), the sum of the weights is 1, and is set dynamically according to actual conditions.

[0088] Calculate the total flow deviation: ΔD′=(0.3×0.632)+(0.2×0.0152)+(0.25×0.175)+0.25×0.25)=0.299.

[0089] Calculate the methane concentration attenuation coefficient: k = 0.08 × (1-0.299) = 0.0561m -1 .

[0090] The final attenuation coefficient k dropped to 0.0561m -1 , indicating that due to the high wind speed, the diffusion path is elongated, and the decay rate of methane concentration slows down (the diffusion range is expanded). In addition, if the wind speed is low or the wind direction deviates greatly from the main diffusion direction, the total flow deviation ΔD' increases, resulting in a smaller decrease in the methane concentration attenuation coefficient k, a smaller diffusion range, and methane concentration is more likely to accumulate in local areas. According to the wind speed range of 1.2≤V≤8.0m / s, if V<3.0m / s, the wind speed deviation ΔD V >0.7, which increases ΔD', causing k to approach the basic attenuation coefficient k0, and the diffusion path is limited. If the wind direction deviates from the main diffusion direction by more than 30°, the wind direction deviation ΔD θ >0.5, ΔD' increases and the diffusion path deflects more seriously. If the temperature is higher than the standard value and the wind speed is high, the air convection is enhanced, ΔD T Depends on T min =280K and T max =320K, if T>310K, then ΔD T >0.5, resulting in a decrease in ΔD', making k lower than the basic attenuation coefficient k0, and expanding the diffusion range. If the temperature is close to the reference value of 298K, then ΔD T <0.2, ΔD' is mainly determined by wind speed and direction, and temperature has little effect on diffusion. In a high humidity environment, the humidity range is 30% ≤ H ≤ 90%. If H> 80%, then ΔD H>0.5, which increases ΔD', causing k to decrease slowly and methane to accumulate near the ground. If the humidity is close to 50%, ΔD H <0.2, humidity has little effect, ΔD' is mainly controlled by wind speed and direction, and the diffusion pattern is similar to that under low humidity conditions. If the terrain is complex, such as a canyon or a densely built-up area, the wind speed and direction fluctuate greatly, and ΔD' may vary significantly in different areas, resulting in uneven k in the local area. If the height of the terrain obstacle is greater than 10m, the wind flow is disturbed, resulting in ΔD in different areas. V and ΔD θ The calculation results are different, and the final range of variation of k increases, which is manifested as rapid diffusion of methane in some areas, while the concentration in other areas remains at a high level for a long time.

[0091] Analyze the methane concentration value from the first grid point corresponding to the potential leakage area to the second grid point occupied by the terrain obstacle space according to the methane concentration attenuation coefficient, and generate a methane concentration distribution map after terrain analysis;

[0092] Analyze the methane concentration value C of the potential leakage area corresponding to the first grid point to the second grid point occupied by the terrain obstacle space corrected , using the formula: C corrected =C new (i,j)×e -k×d ;

[0093] Where d is the distance from the first grid point corresponding to the potential leakage area to the second grid point occupied by the terrain obstacle space, e is the base of the exponential function, which ensures that the decay rate remains smoothly changing, and k represents the methane concentration decay coefficient.

[0094] Substitute the values ​​into the formula: C corrected =9.8×e -0.0561×50 ≈0.593ppm.

[0095] After considering the influence of terrain obstacles, the methane concentration value C of the potential leakage area corresponding to the first grid point to the second grid point occupied by the terrain obstacle space is corrected The methane concentration value is 0.593 ppm, indicating that the obstacle significantly hinders methane diffusion. Add the methane concentration values ​​from the first grid point of the potential leakage area to the second grid point of the terrain obstacle to the diffusion-optimized methane concentration heat map. Use unified visualization software (such as ParaView, Tecplot, or MATLAB) to present the complete terrain-corrected methane concentration distribution, generating a methane concentration distribution map after terrain analysis.

[0096] Accurate three-dimensional terrain data is obtained through drone LiDAR scanning. Combined with the diffusion-optimized methane concentration heat map, a spatial occupancy model of terrain obstacles is constructed to make the gas diffusion analysis more consistent with the actual terrain characteristics. The finite volume method is used to discretize the diffusion path to ensure more precise gas flow calculations, which helps to reduce misjudgments caused by complex terrain. Flow deviations are analyzed in combination with real-time environmental parameters (wind speed, wind direction, temperature, humidity), and the concentration attenuation coefficient is set to make the gas concentration change trend more consistent with the actual diffusion law. The final generated methane concentration distribution map after terrain analysis can more accurately reflect the diffusion situation in the leakage area, providing a more scientific basis for leakage risk assessment and early warning.

[0097] See also Figure 5 , the specific steps for obtaining the diffusion trend prediction results are:

[0098] Identify the diffusion direction of methane around the potential leakage area in the methane concentration distribution map after terrain analysis and generate methane diffusion direction analysis results;

[0099] Concentration data for the first grid point and surrounding grids in the potential leakage area are obtained. GIS spatial analysis tools are used to screen out matching historical wind field conditions from known wind speed and direction data. Combined with the spatial distribution of terrain obstacles, the influence range of the dominant wind direction on methane diffusion is determined. Within a specific time step, changes in methane concentration are tracked through time series data, and the movement trajectory of high-concentration areas with wind direction is observed. Based on drone cruise data, methane concentration values ​​for multiple time periods are extracted, and the concentration distributions at different times are superimposed and compared to determine the directional trend of methane diffusion. Furthermore, representative diffusion directions are selected. Assuming that in a certain scenario, the terrain is valley terrain, the wind speed is about 5m / s, and the wind direction is southwest-west. After a methane leak, its main diffusion direction will be along the low-lying area of ​​the valley to the northeast-east. In open terrain, due to the wider diffusion range of airflow, the diffusion direction of methane concentration may be more uniform. Finally, the obtained diffusion direction information is superimposed on the methane concentration distribution map after the original terrain analysis in the form of arrow marks to ensure that the diffusion trend around the potential leakage area is visualized.

[0100] Based on the results of methane diffusion direction analysis, all records of historical methane leaks are obtained, and the trend of methane concentration changes in the future is predicted by combining turbulent diffusion theory to generate diffusion trend prediction results;

[0101] First, the historical leakage data repository is called to screen historical cases that meet the current terrain, wind speed, wind direction, relative humidity and other conditions, and methane concentration data for the corresponding time period is extracted. Time series analysis is performed on the selected data, and data outliers and sensor failure points are eliminated to ensure data integrity. Then, the screened data is sorted in chronological order to observe the changing trend of methane concentration. Combined with turbulent diffusion theory, ANSYS Fluent or OpenFOAM is used for simulation, and terrain data scanned by LiDAR is imported. Air flow analysis is performed based on the k-ε turbulence model, and boundary conditions are set, including inlet wind speed, wind direction, leakage source concentration, and surface roughness. Historical leakage source location data is imported, and the movement trajectory of the leaked gas is observed through the Lagrangian particle tracking method. ParaView is used for visualization, and the concentration changes at different time steps are compared to screen out the diffusion trend that matches the current environmental conditions. Furthermore, the predicted concentration changes are stored in chronological order to generate diffusion trend prediction results, so that the possible paths of methane diffusion in the future can be clearly presented.

[0102] By identifying the direction of methane diffusion, the diffusion trend in the leak area is more intuitive, helping to accurately determine the gas's movement path. Combining historical leakage data with turbulent diffusion theory, future methane concentration trends are predicted, improving the ability to predict leak evolution. This process reduces the impact of environmental interference on diffusion analysis, improves prediction accuracy, enables inspectors to identify potential high-risk areas in advance, optimizes inspection scheduling, enhances early warning capabilities, and improves the proactive protection level of gas pipeline leak detection.

[0103] See also Figure 6 The specific steps for obtaining the results of gas pipeline leakage hazard identification are as follows:

[0104] Visualize the methane concentration change trend in the future time from the diffusion trend prediction results on the methane concentration distribution map after terrain analysis, identify the concentration accumulation of methane in different potential leakage areas, and obtain a methane concentration evolution display map;

[0105] First, the diffusion trend forecast data is retrieved and the methane concentration values ​​at each time step in the forecast time series are extracted. The data are spatially aligned according to geographic coordinates to ensure that the data matches the original three-dimensional terrain grid. Then, using a GIS visualization tool (such as ArcGIS or QGIS) or fluid simulation visualization software (such as ParaView or Tecplot), the diffusion trend forecast data is loaded and the time step parameters are set so that the visualization interface can display the concentration distribution changes at different times. During the specific display, to ensure the readability of the data, a color gradient mapping scheme is set, marking high-concentration areas in red and low-concentration areas in blue, and transparency control is used to avoid overlapping effects of different concentration intervals. At the same time, to enhance the trend display effect, the wind direction vector field is superimposed during the visualization process, and the wind field data is used to drive the dynamic rendering of the diffusion trajectory. At each time step, the concentration data is smoothed using grid interpolation to reduce the value jumps caused by grid boundaries. Finally, a dynamic heat map animation is generated, allowing users to observe the evolution of methane concentration in the future.

[0106] Based on the methane concentration change trend in the methane concentration evolution display diagram, the risk level of the current potential leakage area is assessed and the gas pipeline leakage hazard identification results are generated;

[0107] First, based on the diffusion trend prediction data, the methane concentration change rate within a specific time window and the spatial distribution range corresponding to the change rate are extracted. According to the existing leakage trend data, the risk assessment indicators are set. For example, if the methane concentration in a certain area increases by more than 50% in one time step, it is marked as a high-risk area. If the concentration increases between 20% and 50%, it is marked as a medium-risk area. The area below 20% but still significantly higher than the background concentration is marked as a low-risk area. Then, the historical leakage data is called to extract historical leakage events at the same spatial location, and the similarity between the current diffusion trend and the historical events is calculated. The time series comparison method is used to analyze the methane concentration increase. Whether the long rate is abnormal. For example, within one time step, the methane concentration at a certain grid point increases from 5ppm to 9.8ppm, an increase of about 96%. The area is marked as a high-risk area, while another grid point increases from 5ppm to 6ppm, an increase of 20%, and is marked as a medium-risk area. Subsequently, combined with the terrain data, it is analyzed whether the terrain obstacles affect the diffusion direction. If the high-risk area is in an area with concentrated wind ducts, the risk level can be appropriately increased. On the contrary, if it is in a low-lying area, the risk level may be reduced. Finally, a gas pipeline leakage hazard identification result is generated, and the high-, medium-, and low-risk areas are marked with different colors and output to the GIS system or visualization platform.

[0108] By visualizing future methane concentration trends, the dynamic evolution of leak areas becomes more intuitive, helping inspectors quickly understand gas diffusion. Leak risk levels are analyzed based on concentration cluster characteristics, improving the accuracy and specificity of hazard identification and reducing the likelihood of false alarms and missed alerts. Ultimately, the leak hazard identification results generated based on dynamic risk assessment provide a scientific basis for inspection decisions, enhancing the early warning capabilities and emergency response efficiency of gas pipeline inspections, and thus more effectively preventing potential safety risks.

[0109] The above are merely preferred embodiments of the present invention and do not limit the present invention in any other form. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. An intelligent identification system for abnormal hidden dangers in pipeline inspection based on drones, characterized by: The system includes: The methane concentration detection module collects UAV flight data to construct a first grid in three-dimensional space, evenly divides the first grid into multiple first grid points, and maps the methane concentration value of each first grid point according to the measured laser absorption spectrum data to generate a methane concentration heat map; The concentration data smoothing module reads all methane concentration values ​​in the methane concentration heat map and calculates the local methane concentration gradient at each first grid point. Based on the magnitude of the local methane concentration gradient, it marks the potential leakage area and sets the diffusion coefficient corresponding to the first grid point to update the methane concentration value, thus generating a diffusion-optimized methane concentration heat map. The virtual diffusion field reconstruction module constructs a second grid point occupied by the terrain obstacle space on the diffusion-optimized methane concentration heat map, sets the methane concentration attenuation coefficient at the second grid point, analyzes the methane concentration values ​​from the first grid point to the second grid point based on the methane concentration attenuation coefficient, and generates a methane concentration distribution map after terrain analysis. The leakage trend prediction module predicts the changing trend of methane around the potential leakage area in the methane concentration distribution map after terrain analysis and generates the diffusion trend prediction result; The leakage risk assessment module identifies the changing trend of methane in the diffusion trend prediction results and visualizes it on the methane concentration distribution map after terrain analysis, generating the gas pipeline leakage hazard identification results.

2. The intelligent identification system for abnormal hidden dangers in pipeline inspection based on drones according to claim 1 is characterized in that: The steps for obtaining the methane concentration heat map are specifically as follows: Collecting drone flight data, including drone location information and flight trajectory data along the gas pipeline area, determining the three-dimensional coordinates of all measurement points, integrating the three-dimensional coordinates of all measurement points to construct a first grid in the three-dimensional space, evenly dividing the first grid into a plurality of first grid points, and generating a first grid division result; Acquiring terrain height data, adjusting the first grid points in the first grid division result to match the current terrain structure, collecting laser absorption spectrum data measured by a methane laser telemeter on a UAV, analyzing the methane concentration in the air based on the laser absorption spectrum data, and generating a methane concentration value; The gaps in the methane concentration values ​​between the measurement points are filled by an interpolation method and mapped to each first grid point in the three-dimensional space to generate a methane concentration heat map.

3. The intelligent identification system for abnormal hidden dangers in pipeline inspection based on drones according to claim 1 is characterized in that: The steps for obtaining the methane concentration heat map after the diffusion optimization are as follows: Reading all methane concentration values ​​in the methane concentration thermodynamic map, calculating the local methane concentration gradient at each first grid point, comparing the local methane concentration gradient with a preset change threshold, marking potential leakage areas where the local methane concentration gradient exceeds the change threshold, and generating a potential leakage area marking result; Based on the local methane concentration gradient of the first grid point corresponding to each potential leakage area in the potential leakage area marking result, the diffusion coefficient of the first grid point corresponding to the potential leakage area is set, and the methane concentration value of the first grid point corresponding to the potential leakage area is updated through the partial differential diffusion equation according to the diffusion coefficient to generate a diffusion-optimized methane concentration thermodynamic map.

4. The intelligent identification system for abnormal hidden dangers in pipeline inspection based on drones according to claim 3 is characterized in that: Through the partial differential diffusion equation, the formula is adopted: Calculate the updated methane concentration value C corresponding to the first grid point (i, j) in the potential leakage area new (i,j); Where C(i, j) is the methane concentration value corresponding to the first grid point (i, j) in the potential leakage area, D(i, j) is the diffusion coefficient set according to the first grid point in the potential leakage area, Δx, Δy are the grid steps along the x and y directions of the first grid point in the potential leakage area, Δt is the time step set according to the grid step, S x is the Laplace term of the methane concentration along the x direction at the first grid point corresponding to the potential leakage area, S y is the Laplace term of the methane concentration along the y direction at the first grid point corresponding to the potential leakage area, and Both factors are dimensionless and represent the proportion of the contribution of diffusion to the concentration update during that time step.

5. The intelligent identification system for abnormal hidden dangers in pipeline inspection based on drones according to claim 1 is characterized in that: The steps for obtaining the methane concentration distribution map after the terrain analysis are specifically as follows: Collecting three-dimensional terrain point cloud data scanned by UAV LiDAR, constructing a second grid point occupied by the terrain obstacle space on the diffusion-optimized methane concentration heat map based on the three-dimensional terrain point cloud data, and discretizing the methane diffusion path around the potential leakage area using the finite volume method with reference to the position and distribution of the second grid point to generate a methane diffusion path processing result; Obtain real-time wind speed, wind direction, temperature, and humidity data for the terrain obstacle area corresponding to the methane diffusion path processing results, analyze the total flow deviation of methane in the terrain obstacle area, and set the methane concentration attenuation coefficient of the second grid point occupied by the obstacle space based on the total flow deviation; The methane concentration values ​​from the first grid point corresponding to the potential leakage area to the second grid point occupied by the terrain obstacle space are analyzed according to the methane concentration attenuation coefficient to generate a methane concentration distribution map after terrain analysis.

6. The intelligent identification system for abnormal hidden dangers in pipeline inspection based on drones according to claim 5 is characterized in that: The analysis of the total flow deviation of methane in the terrain obstacle area adopts the formula: ΔD'=w V ΔD V +in θ ΔD θ +in T ΔD T +in H ΔD H The total flow deviation ΔD' is calculated; Among them, w V 、w θ 、w T 、w H Represent the relative influence weights of wind speed, wind direction, temperature and humidity on methane diffusion, ΔD V is the normalized deviation of the wind speed in the current terrain obstacle area relative to the historical wind speed range in the area, ΔD θ is the degree of deviation between the wind direction in the current terrain obstacle area and the main diffusion direction of methane, ΔD T is the normalized deviation between the current terrain obstacle area temperature and the preset reference temperature, ΔD H It is the normalized deviation between the current humidity in the terrain obstacle area and the preset reference humidity.

7. The intelligent identification system for abnormal hidden dangers in pipeline inspection based on drones according to claim 1 is characterized in that: The steps for obtaining the diffusion trend prediction result are specifically as follows: Identifying the diffusion direction of methane around the potential leakage area in the methane concentration distribution map after the terrain analysis, and generating a methane diffusion direction analysis result; Based on the methane diffusion direction analysis results, all records of historical methane leakage are obtained, and the changing trend of methane concentration in the future is predicted in combination with turbulent diffusion theory to generate a diffusion trend prediction result.

8. The intelligent identification system for abnormal hidden dangers in pipeline inspection based on drones according to claim 1 is characterized in that: The steps for obtaining the gas pipeline leakage hidden danger identification result are specifically as follows: Visually display the methane concentration change trend in the future time in the diffusion trend prediction result on the methane concentration distribution map after the terrain analysis, identify the accumulation of methane concentration in different potential leakage areas, and obtain a methane concentration evolution display map; Based on the methane concentration change trend in the methane concentration evolution display diagram, the risk level of the current potential leakage area is evaluated, and a gas pipeline leakage hazard identification result is generated.

Images