Pipeline patrol system based on Al intelligent algorithm

Abstract

The invention discloses a pipeline patrol system based on an Al intelligent algorithm, and relates to the technical field of pipeline patrol, the pipeline patrol system comprises an information acquisition end for information interaction, a server and an execution patrol end, the server comprises an analysis identification module and an execution control module, the execution patrol end comprises an unmanned aerial vehicle and a pipeline robot, the information acquisition end is used for acquiring environmental information in an oil and gas pipeline area. Inspection and fault detection can be effectively carried out in various environments, the continuity and stability of pipeline inspection are ensured, the unmanned aerial vehicle can efficiently cover the high altitude and the difficult-to-reach area of the pipeline, the pipeline robot can go deep into the underground pipeline for inspection, the requirement for manual inspection is reduced, and the working efficiency is improved. And the coverage range and the meticulous degree of patrol are improved, the patrol efficiency is greatly improved, the fault detection time is greatly shortened, and the patrol frequency and the patrol accuracy are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of pipeline patrol, and in particular to a pipeline patrol system based on an AI intelligent algorithm. Background Art

[0002] Pipelines are an important part of modern industry and infrastructure, and pipeline maintenance is related to national energy security. Pipeline maintenance includes various pipeline systems such as oil, natural gas, water supply, and heating. They play a vital role in daily life and industrial production, so it is very important to ensure the safety and stable operation of pipelines.

[0003] Traditional pipeline patrol methods usually rely on manual inspections, where inspectors walk along the pipeline route or use vehicles for regular inspections. During the inspection process, inspectors mainly rely on visual inspections of the pipeline and its surrounding environment. For example, check whether there are signs of corrosion on the pipeline surface, such as rust spots and pits; check whether the pipeline has leaks, which are manifested as oil stains and water stains on the ground; observe whether there are third-party construction activities around the pipeline, such as new excavation equipment and construction personnel that may threaten the safety of the pipeline. For some parts that are difficult to observe directly, inspectors may use simple tools, such as inspection mirrors, to view hidden locations such as the bottom or side of the pipeline. At the same time, they will also record the time, location, problems found, and other information of the inspection for subsequent analysis and processing. This method that relies on manual inspections first wastes human resources and has low inspection efficiency, and has certain limitations in terms of safety, efficiency, and economy.

[0004] Currently, no effective solution has been proposed for the problems in the related technologies. Summary of the invention

[0005] In view of the problems in the related technology, the present invention proposes a pipeline patrol system based on AI intelligent algorithm to overcome the above-mentioned technical problems existing in the existing related technology.

[0006] The technical solution of the present invention is achieved in this way:

[0007] A pipeline patrol system based on AI intelligent algorithm, comprising: an information collection end for information interaction, a service end and an execution patrol end, the service end comprising: an analysis and identification module and an execution control module, the execution patrol end comprising: a drone and a pipeline robot, wherein;

[0008] The information collection terminal is used to collect environmental information in the oil and gas pipeline area, wherein the environmental information at least includes: a vibration sensor deployed in the oil and gas pipeline area to collect oil and gas pipeline vibration values, a pressure sensor deployed in the oil and gas pipeline area to collect oil and gas pipeline pressure values, a temperature sensor deployed in the oil and gas pipeline area to collect oil and gas pipeline temperature values, a flow sensor deployed in the oil and gas pipeline area to collect oil and gas pipeline flow values, and an acoustic wave sensor deployed in the oil and gas pipeline area to collect oil and gas pipeline acoustic wave values;

[0009] The analysis and identification module is used to obtain environmental information in the oil and gas pipeline area, and when the environmental information in the oil and gas pipeline area is abnormal, control the execution control module to perform inspection tasks, and receive external environmental image information or pipeline internal environmental image information collected by the execution control module for analysis and identification;

[0010] The execution control module is used to send an inspection task execution instruction to the inspection execution terminal when the environmental information in the oil and gas pipeline area is abnormal, wherein the inspection task instruction at least includes: setting an inspection path;

[0011] The drone is equipped with a high-resolution camera and acquires image information of the external environment of the oil and gas pipeline according to a preset inspection path and transmits it to the server;

[0012] The pipeline robot is equipped with a high-resolution camera, and obtains image information of the internal environment of the oil and gas pipeline according to a preset inspection path, and transmits it to the service end.

[0013] Further, identifying abnormal environmental information in the oil and gas pipeline area includes the following steps:

[0014] Setting environmental safety threshold ranges for environmental information in the oil and gas pipeline area, including: setting vibration safety threshold ranges, pressure safety threshold ranges, temperature safety threshold ranges, flow safety threshold ranges, and sound wave safety threshold ranges;

[0015] If the vibration value, pressure value, temperature value, flow value and sound wave value of any one of the environmental information currently collected in the oil and gas pipeline area does not meet the environmental safety threshold range, it means that the environmental information in the current oil and gas pipeline area is in an abnormal state.

[0016] Furthermore, the analysis and identification of the external environment image information includes the following steps:

[0017] Preliminary construction of a scene classification model includes: obtaining an external environment image training set in advance, and labeling the external environment image training set, including: labeling normal scenes and abnormal scenes, wherein the normal operation scene of the pipeline is labeled as a normal scene, and the image data of mechanical construction and human cultivation is labeled as an abnormal scene;

[0018] The convolutional neural network (CNN) in deep learning is used to build a scene classification model, and feature extraction calculations are performed based on the convolutional neural network (CNN), and abnormal scene recognition is performed.

[0019] Furthermore, the abnormal scene identification includes the following steps:

[0020] Calculate the distance between the abnormal scene and the oil and gas pipeline, which includes: obtaining the position coordinates of the object in the normal scene and the position coordinates of the oil and gas pipeline, and calculating the risk distance d through the Euclidean distance, which is expressed as:

[0021]

[0022] Conduct risk assessment, including:

[0023] If the risk distance d is less than the safety threshold, the alarm system is triggered to notify relevant personnel to take measures, such as stopping construction or guiding personnel to leave the area near the pipeline;

[0024] If the risk distance d ≥ safety threshold, record and further observe.

[0025] Furthermore, the analysis and identification of the pipeline internal environment image information includes the following steps:

[0026] Pre-read the internal environment image information of the pipeline and perform image preprocessing, including grayscale, filtering, denoising and normalization;

[0027] Performing feature extraction on the processed image includes the following steps:

[0028] Pre-smooth the image with a Gaussian filter to reduce the impact of noise on edge detection;

[0029] Calculate the gradient magnitude and direction of the image, refine the edge by non-maximum suppression, and the gradient magnitude is expressed as:

[0030] Among them, G x and G y is the gradient of the image in the x and y directions;

[0031] Non-maximum suppression is to determine whether the gradient amplitude of the current pixel is the local maximum in the gradient direction. If not, it is set to 0 to obtain a refined edge.

[0032] Use dual thresholds to detect and connect edges, including: Setting a high threshold T h and low threshold T l , greater than T h The pixel point is determined as the edge, which is less than T l The pixels between the two are determined as non-edges, and the pixels between them are determined as edges if they are connected to the determined edge pixels.

[0033] The extracted edges are classified and identified according to the support vector machine (SVM), and the identification results are obtained.

[0034] Furthermore, obtaining the recognition result comprises the following steps:

[0035] Extract the features of the internal environment image of the pipeline and obtain the feature vector x new Substituting the function into the hyperplane equation, it is expressed as:

[0036] f(x)=w·x new +b;

[0037] Among them, if f(x new )≥0, the extracted feature is judged to belong to the category label, that is, the corresponding pipeline area has defects;

[0038] Among them, if f(x new )<0, it is determined that the extracted feature does not belong to the category label, that is, there is no defect in the corresponding pipeline area.

[0039] Beneficial effects of the present invention:

[0040] The present invention integrates an information collection end, a service end and an execution inspection end, collects environmental information in the oil and gas pipeline area, and controls the execution control module to perform inspection tasks when the environmental information in the oil and gas pipeline area is abnormal. The preset inspection path controls the drone and the pipeline robot to obtain the external environment image information of the oil and gas pipeline and the internal environment image information of the oil and gas pipeline respectively. At the same time, the analysis and recognition module receives the external environment image information or the internal environment image information of the pipeline collected by the execution control module for analysis and recognition, so as to predict potential faults according to real-time data, help the operation to take preventive measures in advance, significantly reduce the risk of sudden failures, and improve the reliability and safety of the overall system. At the same time, the drone and the pipeline robot can cover more oil and gas pipeline areas, can effectively perform inspections and fault detection in a variety of environments, and ensure the continuity and stability of pipeline patrol. The drone can efficiently cover the high altitude and hard-to-reach areas of the pipeline, and the pipeline robot can go deep into the underground pipeline for inspection, which not only reduces the need for manual inspection, but also improves the coverage and meticulousness of the inspection, greatly improves the efficiency of patrol, greatly shortens the time of fault detection, and improves the frequency and accuracy of patrol.

[0041] In addition, based on real-time data and prediction results, maintenance personnel and equipment resources can be reasonably arranged to avoid problems such as excessive or insufficient maintenance, improve the efficiency of maintenance work, and reduce operating costs. Maintenance time can be arranged more accurately, pipeline downtime can be reduced, and the continuity and stability of oil and gas transportation can be ensured. An efficient, accurate and safe pipeline monitoring and maintenance solution is provided, which improves the efficiency and reliability of patrols, strengthens safety, and provides strong support for the sustainable development and safe operation of the oil and gas industry. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0043] Figure 1 is a principle block diagram of a pipeline patrol system based on an AI intelligent algorithm according to an embodiment of the present invention;

[0044] Figure 2 It is a scene schematic diagram of a pipeline patrol system based on AI intelligent algorithm according to an embodiment of the present invention. DETAILED DESCRIPTION

[0045] The following will be combined with the accompanying drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field belong to the scope of protection of the present invention.

[0046] According to an embodiment of the present invention, a pipeline patrol system based on AI intelligent algorithm is provided.

[0047] like Figure 1-Figure 2 As shown, the pipeline patrol system based on the AI ​​intelligent algorithm according to an embodiment of the present invention includes: an information collection terminal 1 for information interaction, a service terminal 2 and an inspection execution terminal 3, the service terminal 2 includes: an analysis and identification module 21 and an execution control module 22, and the inspection execution terminal 3 includes: a drone 31 and a pipeline robot 32, wherein;

[0048] The information collection terminal 1 is used to collect environmental information in the oil and gas pipeline area, wherein the environmental information at least includes: a vibration sensor deployed in the oil and gas pipeline area to collect oil and gas pipeline vibration values, a pressure sensor deployed in the oil and gas pipeline area to collect oil and gas pipeline pressure values, a temperature sensor deployed in the oil and gas pipeline area to collect oil and gas pipeline temperature values, a flow sensor deployed in the oil and gas pipeline area to collect oil and gas pipeline flow values, and an acoustic sensor deployed in the oil and gas pipeline area to collect oil and gas pipeline acoustic values;

[0049] Specifically, during application, the vibration sensor is installed next to the ground pile in the oil and gas pipeline area. When there is vibration near the ground pile, the vibration sensor uploads the location information and alarm information to the server 2, and uses the vibration in the oil and gas pipeline area to determine whether construction or other operations are being carried out in the oil and gas pipeline area.

[0050] Meanwhile, pressure sensors monitor pressure fluctuations inside oil and gas pipelines to identify possible pressure anomalies or leaks. Temperature sensors track temperature changes in oil and gas pipelines to prevent material fatigue or damage caused by overheating or overcooling. Flow sensors measure fluid flow and identify flow anomalies that may indicate blockage or leaks in oil and gas pipelines. Acoustic wave sensors: Use acoustic wave technology to detect cracks or structural defects in oil and gas pipelines.

[0051] The analysis and identification module 21 is used to obtain environmental information in the oil and gas pipeline area, and when the environmental information in the oil and gas pipeline area is abnormal, control the execution control module 22 to perform the inspection task, and receive the external environment image information or the pipeline internal environment image information collected by the execution control module 22 for analysis and identification;

[0052] This technical solution, when the environmental information in the oil and gas pipeline area is abnormal, includes the following steps:

[0053] The environmental safety threshold range is preset for the environmental information in the oil and gas pipeline area, including: setting the vibration safety threshold range, pressure safety threshold range, temperature safety threshold range, flow safety threshold range and sound wave safety threshold range;

[0054] If any vibration value, pressure value, temperature value, flow value and sound wave value in the currently collected environmental information in the oil and gas pipeline area does not meet the environmental safety threshold range, it means that the current environmental information in the oil and gas pipeline area is in an abnormal state.

[0055] The execution control module 22 is used to send an inspection task execution instruction to the inspection execution terminal 3 when the environmental information in the oil and gas pipeline area is abnormal, wherein the inspection task instruction at least includes: setting an inspection path;

[0056] The drone 31 is equipped with a high-resolution camera and acquires image information of the external environment of the oil and gas pipeline according to a preset inspection path and transmits it to the server 2;

[0057] The pipeline robot 32 is equipped with a high-resolution camera and obtains image information of the internal environment of the oil and gas pipeline according to a preset inspection path and transmits it to the server 2;

[0058] In this technical solution, for analyzing and identifying the external environment image information or the pipeline internal environment image information collected by the receiving execution control module 22, the specific steps are as follows:

[0059] The analysis and identification of the external environment image information includes the following steps:

[0060] Preliminary construction of a scene classification model includes: obtaining an external environment image training set in advance, and labeling the external environment image training set, including: labeling normal scenes and abnormal scenes, wherein the normal operation scene of the pipeline is labeled as a normal scene, and the image data of mechanical construction and human cultivation is labeled as an abnormal scene;

[0061] The convolutional neural network (CNN) in deep learning is used to build a scene classification model. Specifically, ResNet-50 is used, which has 50 layers, including multiple convolutional layers, batch normalization layers, and residual connections. During the training process, the input image is processed through each layer of the network through forward propagation. The convolution layer uses the convolution kernel to extract the features of the image. For example, for a 3×3 convolution kernel, a convolution operation is performed at each position of the image to obtain a local feature map. After multiple convolutional layers, the features of the image are gradually abstracted into high-level semantic features.

[0062] Perform feature extraction calculations based on the convolutional neural network (CNN) and identify abnormal scenes;

[0063] Specifically, the input image size is H×W×CH for height, W for width, C for the number of channels, for color images C=3, the convolution kernel size is K×K×C, the number of channels is the same as the input image, the step size is 8, and the padding is p, then the output feature map is:

[0064] Height, expressed as:

[0065] Width, expressed as:

[0066] Specifically, when the technical solution is applied, a 224×224×3 color image is input, a 7×7×3 convolution kernel is used, the step size is 2, and the padding is 3, then the height of the output feature map is expressed as: Width, expressed as: The number of channels is determined by the number of convolution kernels.

[0067] When an abnormal scenario is identified, analysis is performed according to the preset risk elimination strategy, including the following steps:

[0068] Calculate the distance between the abnormal scene and the oil and gas pipeline, which includes: obtaining the position coordinates of objects in the normal scene such as construction machinery, farmland workers and oil and gas pipelines, and calculating the risk distance d through the Euclidean distance, expressed as:

[0069]

[0070] Conduct risk assessment, including:

[0071] If the risk distance d is less than the safety threshold, the alarm system is triggered to notify relevant personnel to take measures, such as stopping construction or guiding personnel to leave the area near the pipeline;

[0072] If the risk distance d ≥ safety threshold, record and further observe.

[0073] When this technical solution is applied, the safety threshold is 10 meters.

[0074] The analysis and identification of the internal environment image information of the pipeline includes the following steps:

[0075] Pre-read the internal environment image information of the pipeline and perform image preprocessing, including grayscale, filtering, denoising and normalization;

[0076] Among them, grayscale includes: Converting an image into a grayscale image can reduce the amount of data while retaining the basic structural information of the image. The calculation formula for grayscale is:

[0077] Gray=0.299R+0.587G+0.114B;

[0078] Among them, R, G, B are the pixel values ​​of the red, green, and blue channels in the color image, respectively, and Gray is the converted grayscale value.

[0079] Among them, filtering denoising includes: using median filtering to replace the grayscale value of the pixel with the median of the grayscale values ​​of the pixel's neighborhood. For a window, sort the 9 pixel values ​​in the window and take the middle value as the new value of the central pixel. Effectively remove isolated noise points such as salt and pepper noise.

[0080] Among them, normalization: normalize the image pixel values ​​to a specific range, taking values ​​[0,1] or [-1,1], expressed as:

[0081] Among them, x is the original pixel value, x normalized is the normalized pixel value, min and max are the minimum and maximum values ​​of the image pixel value respectively.

[0082] Perform feature extraction on the processed image, specifically including the following steps:

[0083] Pre-smooth the image with a Gaussian filter to reduce the impact of noise on edge detection;

[0084] Calculate the gradient magnitude and direction of the image, refine the edge by non-maximum suppression, and the gradient magnitude is expressed as:

[0085] Among them, G x and G y is the gradient of the image in the x and y directions;

[0086] Non-maximum suppression is to determine whether the gradient amplitude of the current pixel is the local maximum in the gradient direction. If not, it is set to 0 to obtain a refined edge.

[0087] Use dual threshold detection and connect edges, including: dual threshold detection is to set the high threshold T h and low threshold T l , greater than T h The pixel point is determined as the edge, which is less than T l The pixels between the two are determined as non-edges, and the pixels between them are determined as edges if they are connected to the determined edge pixels.

[0088] Classify and identify the extracted edges according to the support vector machine SVM, and obtain the identification results;

[0089] In this technical solution, a pipeline internal environment image training data set (x i ,y i ), x iis a feature vector, expressed as: extracted edge or texture, y i is a category label, which means: crack, corrosion, blockage or leakage;

[0090] Calibrate a hyperplane equation, expressed as:

[0091] w·x+b=0;

[0092] Among them, w is the normal vector of the hyperplane, b is the offset, and x is the extracted feature vector. For the binary classification problem, the hyperplane divides the feature space into two regions, corresponding to two different categories.

[0093] The calibration constraint is y i (w·x i +b)≥1, where yi is the category label of xi.

[0094] Specifically, this technical solution, when applied, includes the following steps:

[0095] Extract the features of the internal environment image of the pipeline and obtain the feature vector x new Substituting the function into the hyperplane equation, it is expressed as:

[0096] f(x)=w·x new +b;

[0097] Among them, if f(x new )≥0, the extracted feature is judged to belong to the category label, that is, the corresponding pipeline area has defects;

[0098] Among them, if f(x new )<0, it is determined that the extracted feature does not belong to the category label, that is, there is no defect in the corresponding pipeline area.

[0099] In summary, with the help of the above technical solution of the present invention, the following effects can be achieved:

[0100] The present invention integrates an information collection terminal 1, a service terminal 2 and an inspection execution terminal 3, collects environmental information in the oil and gas pipeline area, and controls the execution control module 22 to perform inspection tasks when the environmental information in the oil and gas pipeline area is abnormal. The preset inspection path controls the drone 31 and the pipeline robot 32 to obtain the external environment image information of the oil and gas pipeline and the internal environment image information of the oil and gas pipeline respectively. At the same time, the analysis and recognition module 21 receives the external environment image information or the internal environment image information of the pipeline collected by the execution control module 22 for analysis and recognition, so as to predict potential faults according to real-time data, help the operator to take preventive measures in advance, significantly reduce the risk of sudden failures, and improve the reliability and safety of the overall system. At the same time, the drone 31 and the pipeline robot 32 can cover more oil and gas pipeline areas, and can effectively perform inspections and fault detection in a variety of environments to ensure the continuity and stability of pipeline patrol. The drone 31 can efficiently cover the high altitude and difficult-to-reach areas of the pipeline, and the pipeline robot 32 can go deep into the underground pipeline for inspection, which not only reduces the need for manual inspection, but also improves the coverage and meticulousness of the inspection, greatly improves the efficiency of patrol, greatly shortens the time of fault detection, and improves the frequency and accuracy of patrol.

[0101] In addition, based on real-time data and prediction results, maintenance personnel and equipment resources can be reasonably arranged to avoid problems such as excessive or insufficient maintenance, improve the efficiency of maintenance work, and reduce operating costs. Maintenance time can be arranged more accurately, pipeline downtime can be reduced, and the continuity and stability of oil and gas transportation can be ensured. An efficient, accurate and safe pipeline monitoring and maintenance solution is provided, which improves the efficiency and reliability of patrols, strengthens safety, and provides strong support for the sustainable development and safe operation of the oil and gas industry.

[0102] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. After considering the disclosure of the specification and the examples, those skilled in the art will easily think of other embodiments of the present disclosure. This application is intended to cover any modification, use or adaptation of the present disclosure, which follows the general principles of the present disclosure and includes common knowledge or customary technical means in the art that are not disclosed in the present disclosure. The specification and the examples are only regarded as exemplary, and the true scope and spirit of the present disclosure are indicated by the claims.

[0103] It should be understood that the present disclosure is not limited to the exact structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims

1. A pipeline patrol system based on AI intelligent algorithm, characterized in that: include: An information collection end (1), a service end (2) and an inspection execution end (3) for information interaction, wherein the service end (2) comprises: an analysis and identification module (21) and an execution control module (22), and the inspection execution end (3) comprises: a drone (31) and a pipeline robot (32), wherein; The information collection terminal (1) is used to collect environmental information in the oil and gas pipeline area, wherein the environmental information at least includes: a vibration sensor deployed in the oil and gas pipeline area to collect oil and gas pipeline vibration values, a pressure sensor deployed in the oil and gas pipeline area to collect oil and gas pipeline pressure values, a temperature sensor deployed in the oil and gas pipeline area to collect oil and gas pipeline temperature values, a flow sensor deployed in the oil and gas pipeline area to collect oil and gas pipeline flow values, and an acoustic wave sensor deployed in the oil and gas pipeline area to collect oil and gas pipeline acoustic wave values; The analysis and identification module (21) is used to obtain environmental information in the oil and gas pipeline area, and when the environmental information in the oil and gas pipeline area is abnormal, control the execution control module (22) to perform inspection tasks, and receive external environmental image information or pipeline internal environmental image information collected by the execution control module (22) for analysis and identification; The execution control module (22) is used to send an execution inspection task instruction to the execution inspection terminal (3) when the environmental information in the oil and gas pipeline area is abnormal, wherein the inspection task instruction at least includes: setting an inspection path; The drone (31) is equipped with a high-resolution camera and acquires image information of the external environment of the oil and gas pipeline according to a preset inspection path and transmits the image information to the server (2); The pipeline robot (32) is equipped with a high-resolution camera and acquires image information of the internal environment of the oil and gas pipeline according to a preset inspection path, and transmits the image information to the service end (2).

2. The pipeline patrol system based on AI intelligent algorithm according to claim 1 is characterized in that: Identifying abnormal environmental information in the oil and gas pipeline area includes the following steps: Setting environmental safety threshold ranges for environmental information in the oil and gas pipeline area, including: setting vibration safety threshold ranges, pressure safety threshold ranges, temperature safety threshold ranges, flow safety threshold ranges, and sound wave safety threshold ranges; If the vibration value, pressure value, temperature value, flow value and sound wave value of any one of the environmental information currently collected in the oil and gas pipeline area does not meet the environmental safety threshold range, it means that the environmental information in the current oil and gas pipeline area is in an abnormal state.

3. The pipeline patrol system based on AI intelligent algorithm according to claim 1 is characterized in that: The analyzing and identifying the external environment image information comprises the following steps: Preliminary construction of a scene classification model includes: obtaining an external environment image training set in advance, and labeling the external environment image training set, including: labeling normal scenes and abnormal scenes, wherein the normal operation scene of the pipeline is labeled as a normal scene, and the image data of mechanical construction and human cultivation is labeled as an abnormal scene; The convolutional neural network (CNN) in deep learning is used to build a scene classification model, and feature extraction calculations are performed based on the convolutional neural network (CNN), and abnormal scene recognition is performed.

4. The pipeline patrol system based on AI intelligent algorithm according to claim 3 is characterized in that: The abnormal scene identification comprises the following steps: Calculate the distance between the abnormal scene and the oil and gas pipeline, which includes: obtaining the position coordinates of the object in the normal scene and the position coordinates of the oil and gas pipeline, and calculating the risk distance d through the Euclidean distance, which is expressed as: Conduct risk assessment, including: If the risk distance d is less than the safety threshold, the alarm system is triggered to notify relevant personnel to take measures, such as stopping construction or guiding personnel to leave the area near the pipeline; If the risk distance d ≥ safety threshold, record and further observe.

5. The pipeline patrol system based on AI intelligent algorithm according to claim 1 or 4 is characterized in that: The analysis and identification of the pipeline internal environment image information includes the following steps: Pre-read the internal environment image information of the pipeline and perform image preprocessing, including grayscale, filtering, denoising and normalization; Performing feature extraction on the processed image includes the following steps: Pre-smooth the image with a Gaussian filter to reduce the impact of noise on edge detection; Calculate the gradient magnitude and direction of the image, refine the edge by non-maximum suppression, and the gradient magnitude is expressed as: Among them, G x and G y is the gradient of the image in the x and y directions; Non-maximum suppression is to determine whether the gradient amplitude of the current pixel is the local maximum in the gradient direction. If not, it is set to 0 to obtain a refined edge. Use dual thresholds to detect and connect edges, including: Setting a high threshold T h and low threshold T l , greater than T h The pixel point is determined as the edge, which is less than T l The pixels between the two are determined as non-edges, and the pixels between them are determined as edges if they are connected to the determined edge pixels. The extracted edges are classified and identified according to the support vector machine (SVM), and the identification results are obtained.

6. The pipeline patrol system based on AI intelligent algorithm according to claim 5 is characterized in that: The step of obtaining the recognition result comprises the following steps: Extract the features of the internal environment image of the pipeline and obtain the feature vector x new Substituting the function into the hyperplane equation, it is expressed as: f(x)=w·x new +b; Among them, if f(x new )≥0, the extracted feature is judged to belong to the category label, that is, the corresponding pipeline area has defects; Among them, if f(x new )<0, it is determined that the extracted feature does not belong to the category label, that is, there is no defect in the corresponding pipeline area.

Images