Unmanned aerial vehicle infrared thermal imaging method and device for cold leakage detection of low-temperature storage tank

Abstract

The invention discloses an unmanned aerial vehicle infrared thermal imaging method and device for cold leakage detection of a low-temperature storage tank, and relates to the field of automatic inspection and detection of storage tanks, and the method comprises the steps: determining the inspection range and detection distance of an unmanned aerial vehicle based on geometric parameters, a cold leakage frequency region and environment parameters of a site; calculating the distance between the shooting points and the vertical coverage height of single-circle flight, further dividing flight elevation layering, and calculating the number of the shooting points of each circle of flight; establishing a three-dimensional model; collecting a visible light image and an infrared thermal imaging image of each shooting point; the method comprises the following steps: performing multi-dimensional correction and temperature image conversion on an acquired infrared thermal imaging image, performing image registration on a temperature image with a temperature scale and a visible light image, performing cold leakage area identification to obtain a cold leakage area, performing quantitative calculation to obtain three-dimensional positioning of a cold leakage position and a cold leakage area, and performing cold leakage area identification on the cold leakage area. And generating a visual detection result, a detection report and a maintenance suggestion. According to the invention, automatic detection and accurate analysis of the cold leakage area of the low-temperature storage tank can be realized.

Description

Technical Field

[0001] This application relates to the field of automated inspection and testing technology for storage tanks, and in particular to an unmanned aerial vehicle (UAV) infrared thermal imaging method and equipment for detecting cold leakage in cryogenic storage tanks. Background Technology

[0002] Cold leakage in large LNG storage tanks is a common type of facility failure in LNG receiving terminals, mainly caused by the settlement of the insulation layer. If a tank leaks, the LNG medium inside will undergo a phase change due to heating, producing a large amount of BOG gas. This leads to an increase in the consumption of cold stream in the downstream recondenser, and may even cause stratification and turbulence of the liquid inside the tank, resulting in overpressure rupture of the inner tank. Therefore, research on cold leakage in large LNG storage tanks is of great significance.

[0003] The following solutions exist in the existing technology: (1) Distributed optical fiber leakage monitoring system for LNG storage tanks. This solution proposes a distributed optical fiber temperature measuring device for oil tanks, which can detect the leakage area. However, it requires opening optical fiber slots and laying optical fibers on the oil tank in advance, which cannot solve the problem of leakage detection in LNG storage tanks without internal optical fiber laying; (2) Adjustment and positioning method of thermal imager for thermal imaging monitoring of liquefied natural gas storage tanks. This solution proposes a method to obtain the temperature information of the outer wall in the height direction of LNG storage tanks, which realizes the height positioning of the leakage location. However, the operation is relatively complicated and cannot obtain the directional information of the leakage location to achieve automation. Detection; (3) A novel LNG storage tank fault diagnosis method. This scheme proposes an infrared thermal imager real-time monitoring system for LNG storage tanks, which can distinguish the fault types. However, the positioning accuracy of this method is not high enough and it cannot quantify the source of cold leakage; (4) An online heat loss measurement technology based on image stitching infrared thermal imaging scanning. This scheme proposes an infrared thermal imaging method based on UAVs. However, only infrared images are collected, which cannot obtain high-definition image information of the fault location. At the same time, no path planning is proposed for the height direction of the tank. In addition, the distance of the tank area is limited, so it is impossible to collect complete information on the location of frequent cold leakage.

[0004] Further analysis of existing technologies reveals that even detection solutions involving UAV infrared thermal imaging still have the following unresolved engineering challenges.

[0005] The route planning lacks adaptability: the existing plan only mentions the route form of "fixed circular flight + vertical flight", without specifying the calculation basis for the distance between shooting points and the elevation layer - for example, for storage tanks of different diameters (such as 50m and 80m), it is impossible to directly determine the number of shooting points; and the influence of obstructions such as pipe corridors and lightning protection strips around the storage tanks is not considered, which can easily create blind spots in the detection.

[0006] Insufficient accuracy in infrared data processing: Temperature calibration only corrects for ambient temperature, without considering the influence of multiple factors such as atmospheric transmittance (significant attenuation of infrared signals in high humidity environments), lens distortion (edge ​​pixel temperature error exceeding ±3℃), and tank surface reflection (difference in emissivity between metal parts and insulation layer), resulting in large deviations in the quantification of cold leakage.

[0007] Cold leakage identification is prone to misjudgment: relying on a single temperature threshold to determine cold leakage, without combining historical cold leakage data of the storage tank (such as the lowest temperature in the same area in the past) and the characteristics of the medium (such as the difference in cold leakage temperature between liquid nitrogen and LNG), often misjudged as cold leakage in areas shaded by pipelines and obscured by equipment, requiring manual secondary inspection, and cannot achieve true automated analysis.

[0008] Therefore, how to achieve automated and high-precision detection and analysis of cold leakage areas in cryogenic storage tanks, while adapting to multi-media storage tanks and complex field environments, and providing technical support for the long-term safe operation of cryogenic storage tanks, has become a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0009] The purpose of this application is to provide an unmanned aerial vehicle (UAV) infrared thermal imaging method and equipment for detecting cold leakage in cryogenic storage tanks, which can realize automated detection and accurate analysis of cold leakage areas in cryogenic storage tanks.

[0010] To achieve the above objectives, this application provides the following solution.

[0011] In a first aspect, this application provides a UAV infrared thermal imaging method for detecting cold leakage in cryogenic storage tanks, the UAV infrared thermal imaging method for detecting cold leakage in cryogenic storage tanks includes the following steps.

[0012] The characteristic parameters of the tested cryogenic storage tank and the environmental parameters of the site are obtained. The characteristic parameters of the tested cryogenic storage tank include: geometric parameters, medium parameters, and historical test data. The geometric parameters include: tank inner diameter and tank wall height. The medium parameters include: medium type, medium standard boiling point, insulation material, and surface emissivity. The historical test data includes: location coordinates of areas with frequent cold leakage, minimum temperature value, and cold leakage area. The environmental parameters of the site include: ambient temperature, relative humidity, atmospheric pressure, wind speed, and atmospheric transmittance.

[0013] Based on the geometric parameters, the areas with frequent cold leakage, and the environmental parameters of the site, the inspection range and detection distance of the UAV are determined.

[0014] The distance between shooting points is calculated based on the detection distance and the horizontal field of view of the infrared lens mounted on the UAV.

[0015] Based on the detection distance and the vertical field of view of the UAV's infrared camera, the vertical coverage height of a single flight circle is calculated.

[0016] Based on the vertical coverage height of the single-circle flight and the height of the tank wall, the flight elevation is divided into layers.

[0017] Based on the flight elevation stratification, the number of shooting points per flight circle is calculated according to the inner diameter of the tank and the distance between shooting points.

[0018] Images of the cryogenic storage tank under test and its environment were collected to create a three-dimensional model.

[0019] Based on the inspection range and detection distance of the UAV and the number of shooting points per flight, visible light images and infrared thermal images of each shooting point are collected; each image is automatically associated with the GPS coordinates of the shooting point, the detection distance, the environmental parameters of the site, and the infrared lens parameters.

[0020] The acquired infrared thermal imaging images are subjected to multi-dimensional correction and temperature image conversion to obtain a temperature image with a temperature scale; the GPS coordinates of the shooting point, the detection time, and environmental parameters are simultaneously marked in the temperature image with the temperature scale.

[0021] The temperature image with the temperature scale is image registered with the visible light image to obtain the registered image.

[0022] The registered image is subjected to cold leakage region identification to obtain the cold leakage region.

[0023] The leakage area is quantitatively calculated to obtain the three-dimensional location of the leakage point and the leakage area.

[0024] Based on the 3D model, the 3D location of the cold leakage point, and the cold leakage area, visualized detection results, detection reports, and maintenance suggestions are generated.

[0025] Secondly, this application provides a drone infrared thermal imaging device for detecting cold leaks in cryogenic storage tanks. The drone infrared thermal imaging device for detecting cold leaks in cryogenic storage tanks is used to implement the drone infrared thermal imaging method for detecting cold leaks in cryogenic storage tanks described in the first aspect. The drone infrared thermal imaging device for detecting cold leaks in cryogenic storage tanks includes: a drone, an image acquisition module, an environmental perception module, a data processing module, and a result visualization module.

[0026] The drone supports automatic route planning and execution, has automatic obstacle avoidance function, and automatically hovers and takes pictures according to the preset route.

[0027] The image acquisition module includes a visible light camera and an infrared thermal imaging lens, which are mounted on the UAV and used to acquire visible light images and infrared thermal imaging images.

[0028] The environmental sensing module is used to collect environmental parameters on site.

[0029] The data processing module is used to acquire the characteristic parameters of the tested cryogenic storage tank and the environmental parameters of the site. The characteristic parameters of the tested cryogenic storage tank include geometric parameters, medium parameters, and historical detection data. The geometric parameters include the tank's inner diameter and tank wall height. The medium parameters include the medium type, standard boiling point, insulation material, and surface emissivity. The historical detection data includes the location coordinates, lowest temperature value, and leakage area of ​​the frequently leaking area. The environmental parameters of the site include ambient temperature, relative humidity, atmospheric pressure, wind speed, and atmospheric transmittance. Based on the geometric parameters, the frequently leaking area, and the environmental parameters of the site, the inspection range and detection distance of the UAV are determined. Based on the detection distance and the horizontal field of view of the infrared lens mounted on the UAV, the distance between shooting points is calculated. Based on the detection distance and the vertical field of view of the UAV's infrared lens, the vertical coverage height of a single-circle flight is calculated. Based on the vertical coverage height of a single-circle flight and the tank wall height, the flight altitude is divided. The process involves layering the flight altitude; calculating the number of shooting points per flight circle based on the tank's inner diameter and the distance between shooting points; acquiring images of the tested cryogenic storage tank and its environment to establish a 3D model; acquiring visible light and infrared thermal images of each shooting point based on the UAV's inspection range, detection distance, and the number of shooting points per flight circle; automatically associating each image with the GPS coordinates of the shooting point, detection distance, on-site environmental parameters, and infrared lens parameters; performing multi-dimensional correction and temperature image conversion on the acquired infrared thermal images to obtain temperature images with temperature scales; synchronously labeling the GPS coordinates of the shooting points, detection time, and environmental parameters in the temperature images with temperature scales; performing image registration between the temperature images with temperature scales and the visible light images to obtain registered images; identifying cold leakage areas in the registered images to obtain cold leakage areas; and performing quantitative calculations on the cold leakage areas to obtain the 3D location and area of ​​the cold leakage.

[0030] The result visualization module is used to generate visualized detection results, detection reports, and maintenance suggestions based on the three-dimensional model, the three-dimensional location of the leakage point, and the leakage area.

[0031] Thirdly, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and capable of running on the processor, wherein the processor executes the computer program to implement the UAV infrared thermal imaging method for detecting cold leakage in cryogenic storage tanks as described in the first aspect.

[0032] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the UAV infrared thermal imaging method for detecting cold leakage in cryogenic storage tanks as described in the first aspect.

[0033] Fifthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the UAV infrared thermal imaging method for detecting cold leakage in cryogenic storage tanks as described in the first aspect.

[0034] Based on the specific embodiments provided in this application, the following technical effects are disclosed.

[0035] This application provides an unmanned aerial vehicle (UAV) infrared thermal imaging method and device for detecting cold leakage in cryogenic storage tanks. The method includes: acquiring characteristic parameters of the cryogenic storage tank under test and environmental parameters of the site; the characteristic parameters of the cryogenic storage tank under test include: geometric parameters, medium parameters, and historical detection data; the geometric parameters include: tank inner diameter and tank wall height; the medium parameters include: medium type, standard boiling point of the medium, insulation layer material, and surface emissivity; the historical detection data includes: location coordinates of frequently leaking areas, minimum temperature value, and leaking area; the environmental parameters of the site include: ambient temperature, relative humidity, atmospheric pressure, wind speed, and atmospheric transmittance; providing a comprehensive and accurate data foundation for precise analysis and calculation in subsequent stages, ensuring that the entire detection process fully considers the characteristics of the storage tank itself and external environmental factors, and improving the reliability of the detection results.

[0036] Based on the geometric parameters, areas prone to cold leakage, and environmental parameters at the site, the inspection range and detection distance of the UAV are determined. This ensures comprehensive coverage of areas where cold leakage may occur, while also allowing for reasonable planning of the detection distance according to actual conditions. This not only guarantees the flight safety of the UAV but also improves the targeting and effectiveness of the inspection.

[0037] Based on the detection distance and the horizontal field of view of the infrared lens on the UAV, the distance between shooting points is calculated. According to the actual detection distance and the observation range characteristics of the lens, the distance between adjacent shooting points can be scientifically determined, ensuring that the images collected during the UAV inspection can completely cover the target area, avoiding detection blind spots, while avoiding excessive collection that leads to data redundancy, making the distribution of image collection points more reasonable, and improving data collection efficiency and quality.

[0038] Based on the detection distance and the vertical field of view of the UAV's infrared lens, the vertical coverage height of a single flight is calculated. Calculating the vertical coverage height of a single flight by using the detection distance and the vertical field of view of the UAV's infrared lens allows for a comprehensive and rational planning of the effective vertical coverage range for each flight, taking into account both the actual distance between the UAV and the target and the vertical observation angle characteristics of the lens itself. This provides a reliable basis for the subsequent accurate division of flight elevation layers, ensuring comprehensive and thorough detection of different height areas of the storage tank throughout the entire flight process.

[0039] Based on the vertical coverage height of a single flight and the tank wall height, flight elevation layers are defined. This method of dividing flight elevation layers according to the vertical coverage height of a single flight and the tank wall height fully considers the effective vertical coverage range of a single UAV flight and the actual height of the storage tank. The entire tank wall height is layered at reasonable intervals, allowing the UAV to systematically inspect different height areas of the storage tank sequentially. This ensures the comprehensiveness and orderliness of the inspection process and avoids missed or duplicate inspections.

[0040] Based on the aforementioned flight elevation stratification, the number of shooting points per flight lap is calculated according to the inner diameter of the tank and the spacing between the shooting points. This fully considers the geometric structure of the storage tank and the rationality of the distribution of image acquisition points, ensuring that the shooting points are set up to fully cover the circumference of the storage tank at that elevation level during each flight lap, while also guaranteeing a scientifically reasonable distance between adjacent shooting points. This ensures that the collected data has sufficient accuracy and completeness, providing high-quality data support for subsequent image analysis and cold leakage detection.

[0041] Images of the tested cryogenic storage tank and its surrounding environment were collected to create a 3D model. This provided an intuitive and three-dimensional model foundation for the subsequent presentation of cold leakage detection results. It allowed operators to gain a comprehensive understanding of the overall condition of the storage tank and its surrounding environment from multiple perspectives. In subsequent analysis of information such as the location and area of ​​cold leakage, the model could be located and evaluated more clearly and accurately, thereby improving the efficiency and accuracy of the detection result analysis.

[0042] Based on the UAV's inspection range and detection distance, as well as the number of shooting points per flight, visible light and infrared thermal images are collected for each shooting point. Each image is automatically associated with the GPS coordinates of the shooting point, detection distance, on-site environmental parameters, and infrared lens parameters. This enriches the additional information of the image and provides more dimensions of data support for subsequent image analysis and interpretation. When performing cold leakage analysis, this associated information can be combined to more accurately determine the cold leakage situation, and it also facilitates data traceability and comparative analysis.

[0043] The acquired infrared thermal imaging images are subjected to multi-dimensional correction and temperature image conversion to obtain a temperature image with a temperature scale. The temperature image with the temperature scale is simultaneously labeled with the GPS coordinates of the shooting point, the detection time, and environmental parameters. The correction eliminates the errors caused by various factors in the image, making the temperature data more accurate and reliable. The labeling information enriches the background data of the image, making it easier for operators to quickly understand the specific conditions of image acquisition. This is beneficial for subsequent scientific and comprehensive analysis of the image temperature data, and thus accurately judging the situation of cold leakage.

[0044] The temperature image with the temperature scale is registered with the visible light image to obtain a registered image. This fully integrates temperature information and the appearance and texture information of the object, allowing operators to determine the potential location of the cold leakage area based on the temperature anomaly, and to more accurately identify the boundary and range of the cold leakage area by using details such as the tank structure in the visible light image. This provides a more intuitive and comprehensive image basis for subsequent analysis and processing of the cold leakage situation.

[0045] The registered images are used to identify areas where cold leakage occurs, thus obtaining the areas where cold leakage may occur. This allows for the rapid and accurate location of potential cold leakage problems in storage tanks from a large amount of image data, providing a clear target for subsequent quantitative analysis and processing of cold leakage, and greatly improving the targeting and efficiency of the detection work.

[0046] The leakage area is quantitatively calculated to obtain the three-dimensional location and area of ​​the leakage. This provides accurate quantitative data for subsequent assessment of the leakage status of the cryogenic storage tank, enabling operators to accurately grasp the specific location and impact range of the leakage, and providing strong data support for formulating scientific and reasonable maintenance strategies.

[0047] Based on the 3D model, the 3D location of the leak, and the leak area, visualized detection results, detection reports, and maintenance suggestions are generated. Complex detection data is presented in an intuitive and easy-to-understand form, providing tank maintenance personnel with comprehensive and systematic information references, helping them to quickly understand the leak situation, formulate scientific and reasonable maintenance plans, effectively improve the pertinence and efficiency of tank maintenance work, and reduce safety risks.

[0048] This application, through close collaboration across multiple steps—from comprehensive data acquisition and scientific testing planning to precise image analysis and result presentation—significantly improves testing efficiency. The use of drones reduces manpower and testing time; it also enhances testing accuracy, with multi-dimensional correction and quantitative calculations making the results more precise; it ensures comprehensive testing, with rationally planned flight and image acquisition covering the entire storage tank; and it provides comprehensive maintenance guidance, with visualized results and maintenance suggestions offering strong support for tank maintenance work, thus providing a solid guarantee for the safe and stable operation of cryogenic storage tanks. Attached Figure Description

[0049] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0050] Figure 1This is a flowchart illustrating an embodiment of an UAV infrared thermal imaging method for detecting cold leakage in a cryogenic storage tank.

[0051] Figure 2 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0052] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0053] The purpose of this application is to retain the core advantages of automation in UAV infrared thermal imaging detection, solve the problems of poor adaptability to flight path planning and numerous blind spots in existing solutions; establish a multi-dimensional infrared data correction model to reduce the impact of environmental, equipment, and tank surface characteristics on temperature measurement errors and improve the accuracy of cold leakage quantification; form cold leakage identification rules that combine historical data and medium characteristics to reduce misjudgments and achieve full automation of the "collection-analysis-judgment" process; and adapt to cryogenic storage tanks of multiple media such as LNG, liquid nitrogen, and liquid oxygen to broaden the coverage of detection scenarios.

[0054] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0055] In one exemplary embodiment, such as Figure 1 As shown, an infrared thermal imaging method for detecting cold leakage in cryogenic storage tanks using a drone is provided. The method includes the following steps.

[0056] S1: Obtain the characteristic parameters of the tested cryogenic storage tank and the environmental parameters of the site; the characteristic parameters of the tested cryogenic storage tank include: geometric parameters, medium parameters, and historical test data; the geometric parameters include: tank inner diameter and tank wall height; the medium parameters include: medium type, medium standard boiling point, insulation layer material, and surface emissivity; the historical test data includes: location coordinates of areas with frequent cold leakage, minimum temperature value, and cold leakage area; the environmental parameters of the site include: ambient temperature, relative humidity, atmospheric pressure, wind speed, and atmospheric transmittance.

[0057] S2: Based on the geometric parameters, the areas with frequent cold leakage, and the environmental parameters of the site, determine the inspection range and detection distance of the UAV.

[0058] S3: Calculate the distance between shooting points based on the detection distance and the horizontal field of view of the infrared lens on the UAV.

[0059] S4: Based on the detection distance and the vertical field of view of the UAV infrared lens, calculate the vertical coverage height of a single-lap flight.

[0060] S5: Based on the vertical coverage height of the single-circle flight and the height of the tank wall, divide the flight elevation into layers.

[0061] S6: Based on the flight elevation stratification, calculate the number of shooting points per flight circle according to the inner diameter of the tank and the distance between shooting points.

[0062] S7: Collect images of the cryogenic storage tank under test and its environment, and establish a three-dimensional model.

[0063] S8: Based on the inspection range and detection distance of the UAV and the number of shooting points per flight, collect visible light images and infrared thermal images of each shooting point; automatically associate each image with the GPS coordinates of the shooting point, detection distance, on-site environmental parameters and infrared lens parameters.

[0064] S9: Perform multi-dimensional correction and temperature image conversion on the acquired infrared thermal imaging image to obtain a temperature image with a temperature scale; the GPS coordinates of the shooting point, the detection time, and environmental parameters are simultaneously marked in the temperature image with the temperature scale.

[0065] S10: Perform image registration between the temperature image with the temperature scale and the visible light image to obtain the registered image.

[0066] S11: Identify the cold leakage region in the registered image to obtain the cold leakage region.

[0067] S12: Perform quantitative calculations on the cold leakage area to obtain the three-dimensional location of the cold leakage position and the cold leakage area.

[0068] S13: Based on the three-dimensional model, the three-dimensional location of the cold leakage point, and the cold leakage area, generate a visual inspection result, inspection report, and maintenance suggestions.

[0069] This application is applicable to the location, quantitative analysis, and operational status assessment of cold leakage in the insulation layer of cryogenic storage tanks containing different media such as LNG, liquid nitrogen, and liquid oxygen. It can realize automated detection and accurate analysis of cold leakage areas in cryogenic storage tanks, providing technical support for the long-term safe operation of cryogenic storage tanks.

[0070] In an exemplary embodiment, step S1 includes: collecting characteristic parameters of the cryogenic storage tank under test.

[0071] Geometric parameters: For vertical storage tanks, record the inner diameter D and the height H of the tank wall (e.g., for LNG full-containment tanks, D=83.6m and H=50.386m); for spherical storage tanks, record the diameter D of the sphere; at the same time, mark fixed features on the surface of the storage tank (e.g., weld inflection points, lightning protection support points, and the edge of the tank top platform) for subsequent image registration.

[0072] Medium parameters: Specify the type of medium in the storage tank (LNG / liquid nitrogen / liquid oxygen), the standard boiling point of the medium (LNG: -163℃, liquid nitrogen: -196℃, liquid oxygen: -183℃), the material of the insulation layer (expanded perlite / glass fiber blanket) and the surface emissivity ε (expanded perlite insulation layer ε=0.92, metal parts surface ε=0.65).

[0073] Historical detection data: Collect the last 3 tank leakage detection reports, extract the location coordinates, minimum temperature value and leakage area of ​​the frequently leaking area to form a historical database for subsequent leakage threshold correction.

[0074] It also includes: collecting environmental parameters at the testing site.

[0075] Real-time environmental data: Ambient temperature T0 (measurement accuracy ±0.1℃), relative humidity RH (measurement accuracy ±1%), atmospheric pressure P (measurement accuracy ±0.1kPa), and wind speed V (measurement accuracy ±0.1m / s) are collected via a portable weather station.

[0076] Infrared propagation influence parameters: Based on the preset UAV detection distance L (initially set to 20-25m) and relative humidity RH, the atmospheric transmittance τ is calculated using the following formula.

[0077] τ=e^(-0.012×L×RH / 100).

[0078] For example, when L=25m and RH=72%, τ=e^(-0.012×25×72 / 100)=e^(-0.216)≈0.806, which is used for atmospheric attenuation correction of subsequent infrared temperature data.

[0079] In an exemplary embodiment, in step S2, the inspection range and detection distance of the drone are determined: based on the geometric parameters (diameter D, height H) obtained in step S1 and the areas where cold leakage frequently occurs (such as the top of the storage tank 10-15m), combined with the distribution of obstacles such as pipe corridors and street lights on site, the inspection range of the drone is determined to cover the entire circumferential tank wall and the areas where cold leakage frequently occurs, while reserving a safety distance of 3-5m for obstacle avoidance.

[0080] Determine the baseline detection distance L: Under normal conditions (wind speed V≤5, relative humidity RH≤85%), L is set to 25m; under high humidity conditions (RH>85%), L is shortened to 20m to reduce atmospheric attenuation of infrared signals; under strong wind conditions (5<V≤7), L remains at 25m, but data integrity is ensured by increasing the number of shooting points.

[0081] In an exemplary embodiment, in step S3, the shooting point spacing is calculated as follows: based on the horizontal field of view β of the infrared lens mounted on the UAV (e.g., β=49.4° if a FLIR Zenmuse XT2 lens is selected) and the detection distance L, the shooting point spacing S is calculated according to the requirement of "adjacent image overlap rate ≥80%", as shown in the following formula.

[0082] S = 2 × L × tan(β / 2) × 0.8.

[0083] Where “0.8” is the overlap coefficient; for example, when L=25m, S=2×25×tan(49.4° / 2)×0.8≈2×25×0.46×0.8≈18.2m.

[0084] In an exemplary embodiment, in steps S4 and S5, the elevation layer is calculated: based on the vertical field of view α of the infrared lens (e.g., α = 40.4°) and the detection distance L, the vertical coverage height h of a single-circle flight is calculated, as shown in the following formula.

[0085] h = 2 × L × tan(α / 2).

[0086] For example, when L=25m, h=2×25×tan(40.4° / 2)≈2×25×0.38≈18.4m; combined with the tank wall height H, the flight elevation is divided into layers—for example, when H=50.386m, it is divided into 3 layers: the first layer is 10-28.4m, the second layer is 28.4-46.8m, and the third layer is 46.8-55m (including the tank top area).

[0087] In an exemplary embodiment, in step S6, the number of shooting points per flight circle is calculated as follows: based on the inner diameter D of the storage tank and the distance S between the shooting points, the number N of shooting points per flight circle is calculated according to the requirement of "no omissions in the entire circumference", as shown in the following formula.

[0088] N = ceil(π × D / S).

[0089] Where “ceil” is the floor function; for example, when D=83.6m and S=18.2m, N=ceil(π×83.6 / 18.2)=ceil(14.4)≈15 revolutions / revolution.

[0090] In one exemplary embodiment, the UAV infrared thermal imaging method for detecting cold leakage in cryogenic storage tanks further includes: full-scene flight path adaptation adjustment.

[0091] Conventional environmental flight routes (V≤5, RH≤85%): The route adopts a combination of "full-circumferential layered circumferential flight + partial vertical supplementary flight" - the full-circumferential circumferential flight is planned according to 3 elevation layers and 15 shooting points per circle, for a total of 45 basic shooting points; the partial vertical supplementary flight targets the pipe gallery and pipeline obstruction areas (such as the pipe gallery above the bottom of the tank 1-3m), and the shooting point spacing is shortened to 10m to ensure that there are no blind spots in the obstruction areas.

[0092] In strong wind environments (5 < V ≤ 7): the distance between shooting points is shortened to S' = S × 0.7 (e.g., 18.2m × 0.7 ≈ 12.7m), and the number of shooting points per lap is increased to N' = ceil(π × D / S') ≈ 21 per lap; at the same time, the drone hovering shooting time is reduced from 10s to 6s, and gimbal stabilization mode is enabled (e.g., the "tripod mode" of DJI M30T) to avoid wind speed affecting image clarity.

[0093] High humidity environment flight path (RH>85%): The detection distance is shortened to L'=20m, and the infrared lens gain is simultaneously increased from the default 1.0 to 1.3 to compensate for the attenuation of infrared signals by the atmosphere; the shooting point spacing is recalculated according to L', S''=2×20×tan(49.4° / 2)×0.8≈14.6m, and the number of shooting points per lap N''=ceil(π×83.6 / 14.6)≈18 / lap.

[0094] In an exemplary embodiment, the UAV infrared thermal imaging method for detecting cold leakage in cryogenic storage tanks further includes: flight path rehearsal and blind zone verification.

[0095] Based on the images of the storage tank and its surrounding environment collected in step S1, import them into 3D modeling software (such as DJI Terra) to create a 3D model of the storage tank and its surrounding environment. The deviation between the model size and the actual object should be controlled within 1-2 cm.

[0096] Import the planned flight path into the 3D model for pre-simulation, and mark areas that may cause blind spots in detection, such as lightning protection zones, ladders, and pipe corridors.

[0097] For blind spots identified in the pre-training, a "slanted supplementary flight path" is added—the drone flies at a 60° angle to the tank wall, with a 5m spacing between shooting points, to ensure that there are no blind spots in areas prone to cold leakage (such as the tank top platform and areas where the insulation layer is prone to settlement).

[0098] In one exemplary embodiment, step S7 specifically includes the following steps.

[0099] 7.1 Image acquisition preparation.

[0100] Drone debugging: Before testing, the drone (such as DJI Matrice M30T) is calibrated with parameters, including RTK positioning calibration (ensuring dynamic positioning accuracy of ±1cm), gimbal level calibration (ensuring shooting angle deviation ≤0.1°), and infrared lens temperature calibration (using a blackbody furnace for calibration, calibration temperature range -180~50℃, calibration error ±0.3℃).

[0101] Equipment mounting: Fix the image acquisition module (including a visible light camera and an infrared lens) to the drone gimbal, ensuring that the optical axes of the visible light camera (4000×3000 resolution) and the infrared lens (1280×1024 resolution) are parallel to avoid image registration deviation.

[0102] 7.2 3D modeling and image acquisition.

[0103] Control the drone to fly at a height of 10-15m above the top of the tank under test, and take pictures of the tank and its surrounding environment from a 45° overhead angle. The picture coverage includes the entire circumference of the tank, the tank top platform and the surrounding environment within 50m (such as pipe corridors and operating platforms).

[0104] During the shooting process, ensure that the overlap rate of adjacent images is ≥70%, and collect no less than 50 regional images in total for subsequent 3D model construction.

[0105] Import the collected area images into 3D modeling software to automatically generate a 3D model of the object under test and its surrounding environment. The model must clearly show structural details such as tank welds, lightning protection strips, and pipe corridors, which will be used for subsequent route optimization and cold leakage location.

[0106] In an exemplary embodiment, in step S8, visible light and infrared thermal imaging images are acquired at each shooting point based on the UAV's flight path, specifically including the following steps.

[0107] 8.1 Synchronous image acquisition.

[0108] The drone flies along a planned, dynamically adaptable flight path. After hovering at each shooting point, it simultaneously collects two sets of image data: ① Visible light image (resolution 4000×3000, focusing on fixed feature points such as tank welds and lightning protection strips, used for image registration and elimination of false cold leakage areas); ② Infrared thermal imaging image (resolution 1280×1024, sampling frequency 5Hz, recording the infrared radiation value of each pixel, used for temperature analysis).

[0109] Data association and storage: Each image is automatically associated with the GPS coordinates of the shooting point (accuracy ±1cm), detection distance L, real-time environmental parameters (T0, RH, V) and infrared lens parameters (gain, focal length). The storage format is XML + image file to ensure data traceability.

[0110] 8.2 Image quality verification.

[0111] For the acquired visible light images, verify the clarity (ensure that feature points such as welds and lightning protection strips are identifiable), and re-acquire blurry images (such as ghosting caused by strong winds).

[0112] For the acquired infrared thermal imaging images, verify the continuity of temperature data (no gaps in the code, no abnormal high / low temperature points). Abnormal data needs to be corrected by re-shooting the drone after it returns to base.

[0113] In an exemplary embodiment, in step S9, the acquired infrared thermal imaging image is subjected to multi-dimensional correction and temperature image conversion to obtain a temperature image with a temperature scale, specifically including the following steps.

[0114] S91: Based on the initial correspondence between infrared radiation value and temperature, the infrared thermal imaging image is initially calibrated to obtain the initial temperature of the infrared thermal imaging image.

[0115] S92: Perform multi-factor temperature correction on the initial temperature of the infrared thermal imaging image to obtain a multi-dimensional corrected infrared thermal imaging image; the multi-factor temperature correction includes: atmospheric transmittance correction, lens distortion correction and surface reflection correction.

[0116] S93: Convert the multi-dimensional corrected infrared thermal imaging image into a temperature image with a temperature scale according to the temperature gradient.

[0117] Specifically, 9.1 Infrared image temperature calibration basic correction.

[0118] Based on the blackbody furnace calibration results before detection, an initial correspondence between infrared radiation values ​​and temperature is established, as shown in the following formula.

[0119] T_initial=K×R+B.

[0120] Where K is the calibration coefficient (e.g., 0.042℃ / count, obtained by fitting the radiation values ​​at different temperature points of the blackbody furnace); R is the radiation value of a pixel in the infrared image (unit: count); and B is the intercept (e.g., -205℃, obtained by linear fitting of the calibration data).

[0121] For example: the infrared radiation value of a certain pixel is R=1200count. Substituting it into the formula, we get T_initial=0.042×1200-205=25.4℃.

[0122] 9.2 Multi-factor temperature correction.

[0123] Atmospheric transmittance correction: Considering the attenuation of infrared signals propagating in the atmosphere, the initial temperature is corrected based on the atmospheric transmittance τ calculated in step S1, as shown in the following formula.

[0124] T_atm = T_initial / τ.

[0125] For example, when τ=0.806, T_atm=25.4 / 0.806≈31.5℃.

[0126] Lens distortion correction: The radial distortion of the infrared lens is corrected using the Brown-Conrady model, as shown in the formula below.

[0127] r_corrected=r×(1+k1×r 2 +k2×r 4 ).

[0128] Where r is the distance from the distorted pixel to the center of the lens (in pixels); k1 and k2 are the lens distortion coefficients (obtained through standard checkerboard calibration, such as k1=-0.012, k2=0.003); after correction, the pixel temperature is recalculated based on r_corrected. For example, when r=500 pixels, r_corrected=500×(1-0.012×500). 2 +0.003×500 4 The image size is approximately 500 × 0.988 = 494 pixels, corresponding to a temperature correction of approximately 31.2℃.

[0129] Surface reflection correction: Considering the influence of the surface emissivity ε of the storage tank on the ambient temperature, the correction formula is as follows.

[0130] T_corrected=T_atm×ε+T0×(1-ε).

[0131] For example: When the expanded perlite insulation layer has an ε=0.92 and T0=28.5℃, T_corrected=31.2×0.92+28.5×(1-0.92)=28.7+2.28=31.0℃ (this is the normal temperature of the tank wall, with no cold leakage).

[0132] 9.3 Temperature Image Conversion.

[0133] The infrared thermal imaging image, after multi-dimensional correction, is converted into a temperature image with a temperature scale according to the temperature gradient. The temperature scale range is preset according to the medium type (LNG tank: -180~40℃, liquid nitrogen tank: -210~40℃) and pseudo-color encoding is used (e.g., ≤-155℃ is red, -155~-150℃ is orange, -150~-145℃ is yellow, and normal temperature is blue).

[0134] The temperature images are simultaneously labeled with the GPS coordinates of the shooting point, the detection time, and environmental parameters (T0, RH) to facilitate subsequent analysis and traceability.

[0135] In an exemplary embodiment, in step S10, the temperature image with the temperature scale is image registered with the visible light image to obtain a registered image, which specifically includes the following steps.

[0136] S101: Based on the visible light image and the marked fixed features of the tank surface, select several non-collinear feature points; the fixed features of the tank surface include: weld inflection points, lightning protection strip support points and the edge of the tank top platform.

[0137] S102: The random sampling consensus algorithm is used to eliminate outliers from the several non-collinear feature points to obtain feature points after outlier elimination.

[0138] S103: Based on the feature points after eliminating outliers, establish a spatial mapping relationship between the infrared thermal imaging image and the visible light image.

[0139] S104: Based on the spatial mapping relationship, the temperature image with the temperature scale is image registered with the visible light image to obtain the registered image.

[0140] Specifically, infrared-visible image registration includes the following steps.

[0141] Feature point selection: Marked fixed feature points of the storage tank (such as weld inflection points, lightning protection belt support points, and tank top platform edges) are used as registration reference points. ≥6 non-collinear feature points are selected for each image pair.

[0142] Registration algorithm: The RANSAC (Random Sample Consensus) algorithm is used to eliminate outliers in feature points and establish a spatial mapping relationship between infrared and visible light images. The registration error is controlled within ≤2 pixels to ensure that the infrared temperature data corresponds accurately with the actual location of the storage tank.

[0143] In an exemplary embodiment, in step S11, the registered image is subjected to cold leakage region identification to obtain the cold leakage region, which specifically includes the following steps.

[0144] S111: Set the initial threshold for judging cold leakage based on the type of medium in the storage tank.

[0145] S112: Based on the ambient temperature at the detection site, the initial threshold for determining cold leakage is corrected to obtain an environmental correction threshold.

[0146] S113: Based on historical cold leakage data of the storage tank, obtain the average of the lowest temperature of the leakage area in the most recent several times, and calculate the historical comparison threshold.

[0147] S114: The minimum value between the environmental correction threshold and the historical comparison threshold is taken as the final judgment threshold.

[0148] S115: Based on the final determination threshold, the registered image is subjected to cold leakage area identification, and false cold leakage areas are excluded to obtain the cold leakage area; the false cold leakage area includes: pipe shadow, equipment obstruction and lens reflection.

[0149] Specifically, the multi-level threshold cold leakage region identification includes the following steps.

[0150] Initial threshold setting: Set the initial cold leakage judgment threshold T1 according to the type of medium in the storage tank - LNG tank T1=-155℃ (12℃ lower than the standard boiling point of the medium), liquid nitrogen tank T1=-188℃ (8℃ lower than the standard boiling point of the medium), liquid oxygen tank T1=-178℃ (5℃ lower than the standard boiling point of the medium).

[0151] Environmental correction threshold: The initial threshold is corrected based on the ambient temperature T0 at the detection site, as shown in the following formula.

[0152] T2 = T1 + 0.05 × (T0 - 25℃).

[0153] Where T2 is the environmental correction threshold; T1 is the initial threshold for determining cold leakage; and T0 is the ambient temperature at the detection site.

[0154] For example: When the LNG tank is at T0=28.5℃, T2=-155+0.05×(28.5-25)=-155+0.175=-154.825℃.

[0155] Historical comparison threshold: If the storage tank has historical cold leakage data, take the average of the lowest temperature of the cold leakage area in the last 3 times, T_history, and calculate the historical comparison threshold: T3=T_history-2℃ (reserve 2℃ error redundancy); where T3 is the historical comparison threshold; T_history is the average of the lowest temperature of the cold leakage area in the last few times.

[0156] Final determination threshold: Take the minimum value between the environmental correction threshold and the historical comparison threshold, that is, T_final=min(T2,T3) -- For example: when T_history=-158℃ for LNG tank, T3=-160℃, then T_final=min(-154.825,-160)=-160℃.

[0157] False cold leakage area elimination: Combine the registered visible light image to eliminate false cold leakage areas such as "pipe shadows, equipment obstruction, lens reflections" - for example, areas that appear as pipe projections in the visible light image are determined to be non-cold leakage areas even if the infrared temperature is lower than T_final.

[0158] In an exemplary embodiment, in step S12, the location of the cold leakage is three-dimensionally located: based on the GPS coordinates (X_GPS, Y_GPS, Z_GPS) of the shooting point and the image registration result, the image coordinates (u, v) of the cold leakage area are converted into the actual three-dimensional coordinates of the tank surface, as shown in the following formula.

[0159] X = X_GPS + (u - u0) × L / f_x.

[0160] Y = Y_GPS + (v - v0) × L / f_y.

[0161] Z = Z_GPS.

[0162] Where (X, Y, Z) are the three-dimensional positioning coordinates of the cold leakage location; (u0, v0) are the coordinates of the center point of the infrared image; f_x and f_y are the horizontal / vertical equivalent focal lengths of the infrared lens (e.g., 280mm); L is the detection distance; for example: when X_GPS=50m, u=640 pixels, u0=640 pixels, L=25m, f_x=280mm, X=50+(640-640)×25 / 0.28=50m.

[0163] Cold leakage area calculation: A combination of "pixel counting method + edge fitting" is adopted - ① Pixel counting: Count the number of pixels N_pixel in the connected region of temperature ≤ T_final in the temperature image; ② Pixel size calculation: Calculate the actual size of a single pixel according to the detection distance L and the field of view of the infrared lens, as shown in the following formula.

[0164] P_size=L×tan(β / 2)×2 / I_width.

[0165] Where P_size is the pixel size; L is the detection distance; β is the horizontal field of view of the infrared lens; and I_width is the width of the infrared thermal imaging image.

[0166] For example: when L=25m, β=49.4°, and I_width=1280 pixels, P_size=25×tan(24.7°)×2 / 1280≈25×0.46×2 / 1280≈0.018m / pixel; ③ Area calculation: Basic area S1=N_pixel×(P_size) 2 Then, a polynomial fitting algorithm is used to optimize the edge of the leakage area (for irregularly shaped leakage areas), and the final leakage area S = S1 × λ (λ is the fitting coefficient, which is adjusted according to the edge smoothness, ranging from 0.97 to 1.03), ensuring that the area calculation accuracy is ≥97%.

[0167] Accuracy verification: Affix a standard marker plate of known size (e.g., 0.5m × 0.5m) to the surface of the storage tank, simultaneously acquire infrared images and calculate the area of ​​the marker plate to verify the calculation error. If the error is >3%, recalibrate the focal length and detection distance of the infrared lens to ensure quantification accuracy.

[0168] In an exemplary embodiment, step S13 outputs the cold leakage detection results and maintenance recommendations, specifically including the following steps.

[0169] 13.1 Visual output of test results.

[0170] Generate a "3D model of the storage tank + annotation map of the cold leakage area": ​​In the established 3D model, the actual 3D coordinates, area and minimum temperature of the cold leakage area are marked, and cold leakage areas of different risk levels are distinguished by different colors (low risk: green, medium risk: yellow, high risk: red).

[0171] Generate an “Infrared-Visible Light Comparison Atlas”: Each area with cold leakage corresponds to one set of comparison images – the left side is the visible light image (marking the boundary of the cold leakage area), and the right side is the temperature image (marking the temperature scale and the lowest temperature of the cold leakage area).

[0172] Generate a "temperature distribution heat map": Integrate the temperature data of the entire storage tank into a heat map, color it according to the temperature gradient, and intuitively display the trend of cold leakage distribution (such as whether there is continuous cold leakage in the 10-15m area at the top of the tank).

[0173] 13.2 Test report generation.

[0174] The report includes: basic detection information (detection time, location, tank number, and medium type), equipment parameters (drone model and infrared lens parameters), environmental parameters (T0, RH, V, τ), flight path planning details (number of shooting points, elevation stratification, and re-flying areas), quantitative results of cold leakage (location coordinates, area, minimum temperature, and risk level of each cold leakage point), and historical data comparison (the trend of cold leakage changes compared to the last three detections).

[0175] Maintenance Recommendations: Based on the leakage area and temperature gradient, tiered maintenance recommendations are provided: ① Low risk (leakage area < 0.5m²) 2 ① ΔT < 5℃: Retest every 3 months to track the trend of cold leakage; ② Medium risk (0.5m 2 ≤Loss area <1m 2 ① 5℃≤ΔT<10℃): Localized filling with insulation material (such as expanded perlite) within 1 month; ③ High risk (leaking area ≥1m²) 2 1. ΔT≥10℃: Shut down within 72 hours to check the settlement of the insulation layer and repair the damaged area; 2. Extremely high risk (leaking area ≥5m²) 2ΔT≥15℃): Emergency shutdown within 24 hours, assessment and replacement of the entire tank's insulation layer.

[0176] In practical applications, the 1.6×10 LNG receiving terminal of the National Pipeline Group in Beihai is used as an example. 5 m 3 The LNG full-containment tank (T-02 tank) is the object under test, and the implementation process of this application is described in detail.

[0177] 1. Basic data collection.

[0178] 1.1 Tank characteristic parameters.

[0179] Geometric parameters: Vertical LNG full-containment tank, inner diameter D=83.6m, tank wall height H=50.386m, tank top platform elevation 33m, cold insulation layer is expanded perlite (thickness 1.0m, ε=0.92); fixed feature points include: vertical welds on the tank wall (12 in total, spaced at 30°), horizontal lightning protection rings (4 rings, with elevations of 27.3m, 31.5m, 35.7m, and 40m respectively), and the edge of the tank top platform.

[0180] Medium parameters: The medium is LNG, with a standard boiling point of -163℃.

[0181] Historical data: In the last three leak detections, leaks were observed in the 12-15m area at the top of the tank, with the lowest temperature T_history = -158℃ and the largest leak area being 0.9m². 2 .

[0182] 1.2 Environmental parameters.

[0183] On the day of the test, the ambient temperature T0=28.5℃, relative humidity RH=72%, atmospheric pressure P=101.2kPa, wind speed V=3.5m / s (normal environment); atmospheric transmittance τ=e^(-0.012×25×72 / 100)=0.806.

[0184] 2. Route planning.

[0185] 2.1 Calculation of core parameters.

[0186] The distance between shooting points S = 2 × 25 × tan(49.4° / 2) × 0.8 ≈ 18.2m.

[0187] Elevation stratification: h = 2 × 25 × tan(40.4° / 2) ≈ 18.4m, divided into 3 layers: 10-28.4m (including 27.3m lightning protection ring), 28.4-46.8m (including 31.5m, 35.7m, and 40m lightning protection rings), and 46.8-55m (including the tank top platform).

[0188] The number of shooting points per lap is N = ceil(π×83.6 / 18.2) = 15 per lap, and the total number of basic shooting points is 15×3 = 45.

[0189] Additional shooting area: 8 additional shooting points (10m apart) above the pipe gallery 1-3m below the tank.

[0190] 2.2 Route execution.

[0191] The drone uses a DJI Matrice M30T. After RTK positioning and calibration, it flies along the planned route, hovering for 10 seconds at each shooting point to simultaneously collect visible light and infrared images. The supplementary flight area of ​​the utility tunnel is shot at a 60° angle to ensure that there are no blind spots above the utility tunnel.

[0192] 3. Image acquisition and data correction.

[0193] 3.1 Image Acquisition.

[0194] A total of 45+8=53 sets of images were collected. The visible light image resolution was 4000×3000, and the infrared thermal image resolution was 1280×1024. Each set of images was associated with GPS coordinates (e.g., GPS for a 10m elevation shooting point: X=100m, Y=80m, Z=10m).

[0195] The infrared lens was calibrated using a blackbody furnace, with K=0.042℃ / count and B=-205℃.

[0196] 3.2 Data correction.

[0197] Initial calibration: For a certain cold leakage area, the pixel count is R=800, and the initial temperature is T_initial=0.042×800-205=31.6-205=-173.4℃.

[0198] Atmospheric correction: T_atm = -173.4 / 0.806 ≈ -215.1℃.

[0199] Distortion correction: r = 400 pixels, r_corrected = 400 × (1 - 0.012 × 400) 2 +0.003×400 4 The image size is approximately 400 × 0.991 = 396.4 pixels, corresponding to a temperature correction of approximately -213.7℃.

[0200] Reflection correction: T_corrected = -213.7 × 0.92 + 28.5 × (1 - 0.92) = -196.6 + 2.28 = -194.3℃ (This is the actual temperature of the cold leakage area).

[0201] 4. Identification and quantification of cold leakage.

[0202] 4.1 Threshold calculation.

[0203] T1 = -155℃ (initial threshold of LNG tank).

[0204] T2=-155+0.05×(28.5-25)=-154.825℃.

[0205] T3 = -158 - 2 = -160℃.

[0206] T_final=min(-154.825,-160)=-160℃.

[0207] 4.2 Cold leakage identification.

[0208] The area with a temperature ≤-160℃ in the infrared image is the 12-15m area at the top of the tank. Combined with the visible light image, it was confirmed that there was no pipe obstruction, and it was determined to be a cold leakage area.

[0209] The reflective area at the 27.3m lightning protection ring on the tank wall was removed (visible light showed a metal lightning protection ring, infrared temperature was -162℃, and it was determined to be a false cold leakage).

[0210] 4.3 Quantitative calculation.

[0211] 3D coordinates: GPS coordinates of the shooting point (X=105m, Y=85m, Z=13.5m), image coordinates of the cold leakage area (u=600, v=500), u0=640, v0=512, f_x=f_y=280mm, L=25m.

[0212] X=105+(600-640)×25 / 0.28≈105-3.57≈101.43m.

[0213] Y=85+(500-512)×25 / 0.28≈85-1.07≈83.93m.

[0214] Z=13.5m.

[0215] The actual coordinates of the cold leakage area are (101.43m, 83.93m, 13.5m).

[0216] Area calculation: Number of pixels in the leaking area N_pixel=3200, P_size=25×tan(24.7°)×2 / 1280≈0.018m / pixel.

[0217] S1 = 3200 × (0.018) 2 ≈1.037m 2 .

[0218] The fitting coefficient λ = 1.002, and the final S = 1.037 × 1.002 ≈ 1.04m 2 .

[0219] 5. Output the results.

[0220] 5.1 Visualization results.

[0221] 3D model annotation: Annotated area of ​​cooling leakage (101.43m, 83.93m, 13.5m), area 1.04m². 2 The lowest temperature is -194.3℃, and the risk level is "high risk" (marked in red). Infrared-visible light comparison chart: The visible light icon on the left marks the boundary of the cold leakage area, and the temperature icon on the right marks the -160℃ threshold line and the lowest temperature.

[0222] Temperature thermogram: The area 12-15m above the tank is shown in red (≤-160℃), and other areas are shown in blue (normal temperature).

[0223] 5.2 Test report and maintenance recommendations.

[0224] Report Conclusion: A high-risk cold leakage point with an area of ​​1.04 m² was found in the 12-15m area above the top of tank T-02. 2 The lowest temperature was -194.3℃.

[0225] Maintenance recommendation: Area with cooling leakage ≥ 1m² 2 It is recommended to stop the machine within 72 hours to check the settlement of the insulation layer in the area and to fill it with expanded perlite insulation material.

[0226] In summary, this application has the following beneficial effects.

[0227] Highly adaptable and covers multiple scenarios: Through dynamic adaptive flight path planning (which can calculate the distance between shooting points and elevation layers), it can adapt to cryogenic storage tanks of different diameters and heights, and at the same time cope with complex environments such as strong winds and high humidity, reducing the detection blind zone rate to ≤1%.

[0228] High detection accuracy and reliable quantification: Multi-dimensional infrared data correction (atmosphere, distortion, reflection) ensures that the temperature measurement error is ≤ ±0.5℃ and the cold leakage area calculation accuracy is ≥97%, which is far superior to existing solutions (temperature error ±2%, area accuracy 90%).

[0229] High degree of automation and low false judgment rate: Multi-level threshold judgment combined with historical data and medium characteristics, along with visible light images, eliminates false cold leakage areas, reducing the false judgment rate of cold leakage to ≤3%, eliminating the need for secondary manual inspection.

[0230] The project is highly practical and easy to implement: Based on mature drones and infrared equipment (such as DJI M30T and FLIR XT2), the software modules support parameter import and manual correction, and on-site personnel can operate it after simple training.

[0231] Data is traceable and supports full lifecycle management: Detection data is linked to GPS, environmental parameters, and historical records, and can be exported to BIM or operation and maintenance systems, providing data support for predicting tank leakage trends and long-term lifecycle management.

[0232] In an exemplary embodiment, a UAV infrared thermal imaging device for detecting cold leaks in cryogenic storage tanks is provided. The UAV infrared thermal imaging device for detecting cold leaks in cryogenic storage tanks is used in the aforementioned UAV infrared thermal imaging method for detecting cold leaks in cryogenic storage tanks. The UAV infrared thermal imaging device for detecting cold leaks in cryogenic storage tanks includes: a UAV, an image acquisition module, an environmental perception module, a data processing module, and a result visualization module.

[0233] The drone supports automatic route planning and execution, has automatic obstacle avoidance function, and automatically hovers and takes pictures according to the preset route.

[0234] The image acquisition module includes a visible light camera and an infrared thermal imaging lens, which are mounted on the UAV and used to acquire visible light images and infrared thermal imaging images.

[0235] The environmental sensing module is used to collect environmental parameters on site.

[0236] The data processing module is used to acquire the characteristic parameters of the tested cryogenic storage tank and the environmental parameters of the site. The characteristic parameters of the tested cryogenic storage tank include geometric parameters, medium parameters, and historical detection data. The geometric parameters include the tank's inner diameter and tank wall height. The medium parameters include the medium type, standard boiling point, insulation material, and surface emissivity. The historical detection data includes the location coordinates, lowest temperature value, and leakage area of ​​the frequently leaking area. The environmental parameters of the site include ambient temperature, relative humidity, atmospheric pressure, wind speed, and atmospheric transmittance. Based on the geometric parameters, the frequently leaking area, and the environmental parameters of the site, the inspection range and detection distance of the UAV are determined. Based on the detection distance and the horizontal field of view of the infrared lens mounted on the UAV, the distance between shooting points is calculated. Based on the detection distance and the vertical field of view of the UAV's infrared lens, the vertical coverage height of a single-circle flight is calculated. Based on the vertical coverage height of a single-circle flight and the tank wall height, the flight altitude is divided. The process involves layering the flight altitude; calculating the number of shooting points per flight circle based on the tank's inner diameter and the distance between shooting points; acquiring images of the tested cryogenic storage tank and its environment to establish a 3D model; acquiring visible light and infrared thermal images of each shooting point based on the UAV's inspection range, detection distance, and the number of shooting points per flight circle; automatically associating each image with the GPS coordinates of the shooting point, detection distance, on-site environmental parameters, and infrared lens parameters; performing multi-dimensional correction and temperature image conversion on the acquired infrared thermal images to obtain temperature images with temperature scales; synchronously labeling the GPS coordinates of the shooting points, detection time, and environmental parameters in the temperature images with temperature scales; performing image registration between the temperature images with temperature scales and the visible light images to obtain registered images; identifying cold leakage areas in the registered images to obtain cold leakage areas; and performing quantitative calculations on the cold leakage areas to obtain the 3D location and area of ​​the cold leakage.

[0237] The result visualization module is used to generate visualized detection results, detection reports, and maintenance suggestions based on the three-dimensional model, the three-dimensional location of the leakage point, and the leakage area.

[0238] Specifically, the characteristics of drones are as follows.

[0239] Model selection: DJI Matrice M30T drone, with a maximum takeoff weight of 4.1kg and a maximum flight time of 41min, meets the requirements for long-haul testing.

[0240] Positioning System: Equipped with an RTK differential positioning system, supporting BeiDou + GPS dual-mode positioning, with static positioning accuracy of ±1mm and dynamic positioning accuracy of ±1cm, ensuring accurate coordinates of the shooting point.

[0241] Anti-interference capability: wind resistance up to level 7, anti-frost coating added to the body (suitable for low temperature environments of -40~50℃), and heating element added to the gimbal (to maintain lens temperature >5℃ in low temperature environments, avoiding lens frost from affecting image formation).

[0242] Flight control: Supports automatic flight path planning and execution, and has automatic obstacle avoidance function (infrared + visual dual obstacle avoidance). It can automatically hover and shoot according to the preset flight path, with a shooting point deviation of ≤±5cm.

[0243] The structure of the image acquisition module is shown below.

[0244] Visible light camera: It adopts a Sony IMX586 sensor with a resolution of 4000×3000, a focal length of 24mm, and an aperture of F2.8. It supports autofocus and exposure compensation to ensure clear imaging of feature points such as tank welds and lightning protection strips.

[0245] Infrared thermal imaging lens: FLIR Zenmuse XT2 dual-light thermal imaging camera with a resolution of 1280×1024, a spectral range of 7.5-13μm, a temperature measurement range of -60~550℃, a NETD (noise equivalent temperature difference) of ≤0.03℃@30℃, and a temperature measurement error of ±2% (after multi-dimensional correction, the error is ≤±0.5℃).

[0246] Synchronous acquisition control: The visible light camera and infrared lens are triggered synchronously through the pan-tilt unit, with an acquisition time difference of ≤10ms, ensuring that the spatial positions of the two images at the same shooting point are accurately corresponded.

[0247] The structure of the environment perception module is shown below.

[0248] Sensor configuration: Integrated temperature and humidity sensor (Sensirion SHT31, measurement range -40~125℃ / 0~100%RH, accuracy ±0.1℃ / ±1%RH), atmospheric pressure sensor (Bosch BMP390, measurement range 300~1250hPa, accuracy ±0.1kPa), and wind speed sensor (Gill WindSonic, measurement range 0~60m / s, accuracy ±0.1m / s).

[0249] Data transmission: Sensor data is transmitted in real time to the data processing module via a 4G / 5G module, with a sampling frequency of 1Hz, for dynamic adjustment of flight paths and infrared data correction.

[0250] The structure of the data processing module is shown below.

[0251] Hardware configuration: Industrial control computer (CPU: Intel Core i7-13700H, 32GB DDR5 memory, 2TB SSD, equipped with NVIDIA RTX 4060 GPU, supporting parallel computing).

[0252] The software functions are as follows.

[0253] Flight route planning submodule: Supports importing 3D models of storage tanks, automatically calculates the spacing between shooting points, elevation layers, and the number of shooting points per loop, and allows manual adjustment of flight route parameters in complex environments (such as increasing the number of shooting points in strong winds).

[0254] Data correction submodule: Built-in multi-dimensional correction algorithm of "initial calibration → atmospheric correction → distortion correction → reflection correction", supports importing blackbody furnace calibration data and lens distortion coefficients.

[0255] Image registration submodule: Integrates RANSAC algorithm to automatically identify fixed feature points of storage tank and complete infrared-visible image registration, with registration error displayed in real time.

[0256] The cold leakage analysis submodule enables multi-level threshold determination (automatic calculation of T1, T2, T3, T_final), pseudo-region exclusion (combined with visible light images), cold leakage quantification calculation (location, area), and supports manual correction of cold leakage region boundaries.

[0257] The structure of the results visualization module is shown below.

[0258] Output formats: Supports generating PDF format inspection reports, 3D model annotation files (which can be imported into AutoCAD and BIM software), infrared-visible light comparison atlases, and temperature distribution heat maps.

[0259] Interactive features: Supports clicking on the leakage area to view detailed parameters (location coordinates, area, temperature, risk level), supports comparison of historical detection data (leakage change trend chart of the same area at different periods), and can export data to the storage tank operation and maintenance management system.

[0260] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 2As shown, the computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs in the non-volatile storage media to run. The database stores characteristic parameters of the tested cryogenic storage tank and on-site environmental parameters. The I / O interfaces are used for information exchange between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a UAV infrared thermal imaging method for detecting cold leakage in cryogenic storage tanks.

[0261] Those skilled in the art will understand that Figure 2 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0262] In one exemplary embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the above-described method embodiments.

[0263] In one exemplary embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the above-described method embodiments.

[0264] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the above-described method embodiments.

[0265] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0266] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0267] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0268] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0269] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for detecting cold leakage of a low-temperature storage tank by using an unmanned aerial vehicle (UAV) infrared thermal imaging, characterized in that, The unmanned aerial vehicle infrared thermal imaging method for detecting cold leakage of the low-temperature storage tank comprises the following steps: Obtain the characteristic parameters of the measured low-temperature storage tank and the environmental parameters of the site; the characteristic parameters of the measured low-temperature storage tank include geometric parameters, medium parameters and historical detection data; the geometric parameters include the inner diameter of the tank body and the height of the tank wall; the medium parameters include the medium type, the medium standard boiling point, the cold insulation layer material and the surface emissivity; the historical detection data include the position coordinates of the cold leakage frequent area, the minimum temperature value and the cold leakage area; the environmental parameters of the site include the environmental temperature, the relative humidity, the atmospheric pressure, the wind speed and the atmospheric transmittance; Determine the unmanned aerial vehicle inspection range and the detection distance based on the geometric parameters, the cold leakage frequent area and the environmental parameters of the site; Calculate the shooting point spacing based on the detection distance and the horizontal field of view angle of the infrared lens carried by the unmanned aerial vehicle; Calculate the vertical coverage height of single circle flight based on the detection distance and the vertical field of view angle of the infrared lens of the unmanned aerial vehicle; Divide the flight elevation layer based on the vertical coverage height of single circle flight and the height of the tank wall; Calculate the number of shooting points of each circle flight based on the inner diameter of the tank body and the shooting point spacing according to the flight elevation layer; Collect the images of the measured low-temperature storage tank and the environment, and establish a three-dimensional model; Collect the visible light image and the infrared thermal imaging image of each shooting point based on the unmanned aerial vehicle inspection range and the detection distance and the number of shooting points of each circle flight; each image is automatically associated with the GPS coordinates of the shooting point, the detection distance, the environmental parameters of the site and the infrared lens parameters; Perform multi-dimensional correction and temperature image conversion on the collected infrared thermal imaging images to obtain a temperature image with a temperature scale; the GPS coordinates of the shooting point, the detection time and the environmental parameters are simultaneously labeled in the temperature image with the temperature scale; Perform image registration on the temperature image with the temperature scale and the visible light image to obtain a registered image; Identify the cold leakage area from the registered image to obtain the cold leakage area; Quantitatively calculate the cold leakage area to obtain the three-dimensional positioning of the cold leakage position and the cold leakage area; Generate a visual detection result, a detection report and a maintenance suggestion based on the three-dimensional model, the three-dimensional positioning of the cold leakage position and the cold leakage area.

2. The unmanned aerial vehicle infrared thermography method for detecting cryogenic tank leaks according to claim 1, wherein, Perform multi-dimensional correction and temperature image conversion on the collected infrared thermal imaging images to obtain a temperature image with a temperature scale, which specifically comprises the following steps: Based on the initial correspondence between the infrared radiation value and the temperature, the initial temperature of the infrared thermal imaging image is obtained through initial calibration; Perform multi-factor temperature correction on the initial temperature of the infrared thermal imaging image to obtain the infrared thermal imaging image after multi-dimensional correction; the multi-factor temperature correction includes atmospheric transmittance correction, lens distortion correction and surface reflection correction; The calculation formula of the atmospheric transmittance correction is as follows: T_atm=T_initial / τ; The calculation formula of the lens distortion correction is as follows: r_corrected=r×(1+k1×r 2 +k2×r 4 ); The calculation formula of the surface reflection correction is as follows: T_corrected=T_atm×ε+T0×(1-ε); Wherein, T_atm is the corrected atmospheric transmittance; T_initial is the initial temperature of the infrared thermal imaging image; τ is the calculated atmospheric transmittance; r_corrected is the corrected radial distortion of the infrared lens; r is the distance from the distorted pixel point to the center of the lens; k1, k2 are the lens distortion coefficients; T_corrected is the corrected surface reflectivity; ε is the surface emissivity; to detect the field environment temperature; The multi-dimension corrected infrared thermal imaging image is converted into a temperature image with a temperature scale according to a temperature gradient.

3. The unmanned aerial vehicle infrared thermography method for detecting cryogenic tank leaks according to claim 1, wherein, The temperature image with the temperature scale is image-matched with the visible light image to obtain a matched image, specifically including: Based on the visible light image and the marked fixed features on the surface of the storage tank, a plurality of non-collinear feature points are selected; the fixed features on the surface of the storage tank include: weld inflection points, lightning belt support points and tank top platform edges; An abnormal value elimination is performed on the plurality of non-collinear feature points by using a random sample consensus algorithm to obtain feature points after elimination of abnormal values; Based on the feature points after elimination of abnormal values, a spatial mapping relationship between the infrared thermal imaging image and the visible light image is established; Based on the spatial mapping relationship, the temperature image with the temperature scale is image-matched with the visible light image to obtain a matched image.

4. The unmanned aerial vehicle infrared thermography method for detecting cryogenic tank leaks according to claim 1, wherein, A cold leakage area is identified from the matched image, specifically including: An initial cold leakage judgment threshold is set according to the type of medium in the storage tank; The initial cold leakage judgment threshold is corrected in combination with the ambient temperature of the detection site to obtain an ambient correction threshold; the calculation formula of the ambient correction threshold is: T2 = T1 + 0.05 × (T0 - 25℃); wherein, T2 is the ambient correction threshold; T1 is the initial cold leakage judgment threshold; and T0 is the ambient temperature of the detection site; Based on historical cold leakage data of the storage tank, an average value of the lowest temperature of the cold leakage area in the last several times is obtained, and a historical comparison threshold is calculated; the calculation formula of the historical comparison threshold is: T3 = T_history - 2℃; wherein, T3 is the historical comparison threshold; and T_history is the average value of the lowest temperature of the cold leakage area in the last several times; The minimum value of the ambient correction threshold and the historical comparison threshold is taken as a final judgment threshold; Based on the final judgment threshold, the cold leakage area is identified from the matched image, and a pseudo cold leakage area is excluded to obtain the cold leakage area; the pseudo cold leakage area includes: pipeline shadow, equipment obstruction and lens reflection.

5. The drone infrared thermography method for detecting cryogenic tank leaks according to claim 1, wherein, The calculation formula of three-dimensional positioning of the cold leakage position is: X = X_GPS + (u-u0) × L / f_x; Y = Y_GPS + (v-v0) × L / f_y; Z = Z_GPS; wherein, (u, v) is the image coordinate of the cold leakage area; (u0, v0) is the center point coordinate of the infrared thermal imaging image; f_x is the horizontal equivalent focal length of the infrared lens; f_y is the vertical equivalent focal length of the infrared lens; L is the detection distance; (X_GPS, Y_GPS, Z_GPS) is the GPS coordinate of the shooting point; and (X, Y, Z) is the three-dimensional positioning coordinate of the cold leakage position.

6. The drone infrared thermography method for detecting cryogenic tank leaks according to claim 1, wherein, The calculation formula of the cold leakage area is: P_size = L × tan(β / 2) × 2 / I_width; S1 = N_pixel x (P_size) 2 ; S = S1 × λ; wherein, P_size is the pixel size; L is the detection distance; β is the horizontal field of view angle of the infrared lens; I_width is the width of the infrared thermal imaging image; S1 is the basic area; N_pixel is the number of pixels; λ is the fitting coefficient; and S is the cold leakage area.

7. A UAV infrared thermal imaging device for low-temperature storage tank cold leakage detection, the UAV infrared thermal imaging device for low-temperature storage tank cold leakage detection is used to realize the low-temperature storage tank cold leakage detection UAV infrared thermal imaging method described in any one of claims 1-6, characterized in that, The unmanned aerial vehicle infrared thermal imaging device for low-temperature tank cold leakage detection comprises an unmanned aerial vehicle, an image acquisition module, an environment perception module, a data processing module and a result visualization module. The unmanned aerial vehicle supports automatic route planning and execution and has an automatic obstacle avoidance function and can automatically hover and shoot according to a preset route. The image acquisition module comprises a visible light camera and an infrared thermal imaging lens and is arranged on the unmanned aerial vehicle to acquire visible light images and infrared thermal imaging images. The environment perception module is used to acquire environmental parameters of a site. The data processing module is used to acquire characteristic parameters of a measured low-temperature tank and environmental parameters of a site. The characteristic parameters of the measured low-temperature tank comprise geometric parameters, medium parameters and historical detection data. The geometric parameters comprise an inner diameter of a tank body and a tank wall height. The medium parameters comprise a medium type, a medium standard boiling point, a cold insulation layer material and a surface emissivity. The historical detection data comprise position coordinates of a cold leakage frequent area, a minimum temperature value and a cold leakage area. The environmental parameters of the site comprise an environmental temperature, a relative humidity, an atmospheric pressure, a wind speed and an atmospheric transmittance. Based on the geometric parameters, the cold leakage frequent area and the environmental parameters of the site, a UAV inspection range and a detection distance are determined. Based on the detection distance and a horizontal field of view angle of an infrared lens carried by the unmanned aerial vehicle, a shooting point spacing is calculated. Based on the detection distance and a vertical field of view angle of the infrared lens of the unmanned aerial vehicle, a vertical coverage height of a single circle flight is calculated. Based on the vertical coverage height of the single circle flight and the tank wall height, a flight elevation layering is divided. Based on the flight elevation layering, a shooting point number of each circle flight is calculated according to the inner diameter of the tank body and the shooting point spacing. Images of the measured low-temperature tank and the environment are acquired to establish a three-dimensional model. Based on the UAV inspection range and the detection distance and the shooting point number of each circle flight, visible light images and infrared thermal imaging images of each shooting point are acquired. Each image is automatically associated with GPS coordinates of a shooting point, a detection distance, environmental parameters of a site and infrared lens parameters. The acquired infrared thermal imaging images are subjected to multi-dimensional correction and temperature image conversion to obtain temperature images with a temperature scale. The temperature images with the temperature scale are simultaneously labeled with GPS coordinates of a shooting point, a detection time and environmental parameters. The temperature images with the temperature scale are subjected to image registration with the visible light images to obtain registered images. The registered images are subjected to cold leakage area identification to obtain a cold leakage area. The cold leakage area is subjected to quantitative calculation to obtain a cold leakage position three-dimensional positioning and a cold leakage area. The result visualization module is used to generate a visual detection result, a detection report and a maintenance suggestion based on the three-dimensional model, the cold leakage position three-dimensional positioning and the cold leakage area.

8. A computer device comprising: The memory, the processor and the computer program stored in the memory and capable of running on the processor are characterized in that the processor executes the computer program to implement the unmanned aerial vehicle infrared thermal imaging method for low-temperature tank cold leakage detection according to any one of claims 1-6.

9. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the unmanned aerial vehicle infrared thermal imaging method for detecting cold leakage of the low-temperature storage tank according to any one of claims 1-6.

10. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to implement the unmanned aerial vehicle infrared thermal imaging method for detecting cold leakage of the low-temperature storage tank according to any one of claims 1-6.

Images