A pipeline monitoring and maintenance method based on drones

Abstract

The invention discloses a pipeline monitoring and maintenance method based on unmanned aerial vehicles (UAVs), and relates to the technical field of pipeline monitoring. The method comprises the following steps: obtaining structural data of a pipeline to be monitored and surrounding network signal strength data, calculating structural similarity and network similarity between the pipeline to be monitored and each pipeline in a monitored pipeline set, and screening out a target monitoring pipeline set and a corresponding UAV set with similar structures and network conditions; calculating a priority index of each target UAV through the remaining power and image upload quality characteristics; selecting a UAV with the highest priority index to perform a monitoring task, and judging whether the pipeline needs maintenance; in this way, for a pipeline that is firstly introduced with UAVs for pipeline monitoring, a suitable UAV can be selected according to actual conditions to monitor the pipeline to be monitored, so that the monitoring data is relatively comprehensive and has high accuracy, and minute cracks or leakage points in the pipeline can be discovered in time, and the pipeline can be maintained in time, thereby reducing the possibility of risks in the pipeline.

Description

Technical Field

[0001] The present invention relates to the technical field of pipeline monitoring, and in particular to a pipeline monitoring and maintenance method based on an unmanned aerial vehicle. Background Art

[0002] UAV-based pipeline monitoring and maintenance refers to the use of drones equipped with high-precision sensors, high-definition cameras and thermal imaging equipment to conduct aerial inspections of pipelines to detect potential faults such as leaks, corrosion, and cracks. The images of the pipelines transmitted in real time by drones are analyzed to determine the operating status of the pipelines and accurately locate the problem areas, thereby improving monitoring efficiency, reducing the risks of manual inspections, and ensuring the safe and stable operation of the pipelines.

[0003] However, in actual pipeline monitoring, for pipelines that are first introduced to monitor with drones, inaccurate selection of drones may result in incomplete or insufficiently accurate drone monitoring data, making it difficult to detect minor cracks or leaks in the pipeline in a timely manner, making it impossible to maintain the pipeline in a timely manner, increasing the possibility of pipeline risks. Summary of the invention

[0004] The purpose of the present invention is to solve the above-mentioned problems and provide a pipeline monitoring and maintenance method based on drones.

[0005] The present invention proposes a pipeline monitoring and maintenance method based on a drone, the method comprising:

[0006] Obtain the self-structure data of the pipeline to be monitored, and record all pipelines that have been monitored by the drone as a monitored pipeline set, and calculate the structural similarity between the pipeline to be monitored and each monitored pipeline according to the self-structure data of the pipeline to be monitored and the self-structure data of each monitored pipeline in the monitored pipeline set;

[0007] Acquire network signal strength data around the pipeline to be monitored, and calculate the network similarity between the pipeline to be monitored and each monitored pipeline according to the network signal strength data of the pipeline to be monitored and the network signal strength data of each monitored pipeline in the monitored pipeline set;

[0008] The monitored pipeline set is screened according to the structural similarity and network similarity to obtain the target monitoring pipeline set, and the drone corresponding to each monitoring pipeline in the target monitoring pipeline set is taken as the target drone to obtain the target drone set;

[0009] Obtain the remaining power of each drone in the target drone set and the image upload quality characteristics when monitoring the corresponding monitored pipeline to obtain the priority selection index of the corresponding drone;

[0010] The drone with the largest index in the target drone set is selected as the final drone to perform flight monitoring on the pipeline to be monitored, and whether the pipeline to be monitored needs maintenance is determined based on the monitoring data.

[0011] Optionally, calculating the structural similarity between the pipeline to be monitored and each monitored pipeline according to the self-structure data of the pipeline to be monitored and the self-structure data of each monitored pipeline in the monitored pipeline set includes:

[0012] For the pipeline to be monitored and each monitored pipeline in the monitored pipeline set, preprocessing the self-structure data of the pipeline to be monitored and the self-structure data of the corresponding monitored pipeline;

[0013] Construct a distance matrix, in which each element is the distance between the self-structure data point of the pipeline to be monitored and the self-structure data point of the monitored pipeline;

[0014] The shortest path in the distance matrix is ​​calculated using a dynamic programming algorithm to find the best match between the self-structure data sequence of the pipeline to be monitored and the self-structure data sequence of the monitored pipeline;

[0015] According to the calculated shortest path, the self-structure data sequence of the pipeline to be monitored and the self-structure data sequence of the monitored pipeline are aligned;

[0016] According to the aligned self-structure data sequence of the pipeline to be monitored and the self-structure data sequence of the monitored pipeline, the similarity between them is calculated, and the calculated similarity is used as the structural similarity between the corresponding monitored pipeline and the pipeline to be monitored, so as to obtain the structural similarity of the pipeline to be monitored and each monitored pipeline in the monitored pipeline set.

[0017] Optionally, calculating the network similarity of the pipeline to be monitored and each monitored pipeline according to the network signal strength data of the pipeline to be monitored and the network signal strength data of each monitored pipeline in the monitored pipeline set includes:

[0018] Continuously acquiring network signal strength data of the pipeline to be monitored within a preset time period at multiple times to obtain a network signal strength sequence of the pipeline to be monitored;

[0019] Calculate the mean and standard deviation of the network signal strength sequence of the pipeline to be monitored as the first mean and first standard deviation;

[0020] Continuously acquiring network signal strength data of each monitored pipeline in the monitored pipeline set within a preset time period at multiple times to obtain a network signal strength sequence of the pipeline to be monitored; calculating the mean and standard deviation of the network signal strength sequence of the monitored pipeline as a second mean and a second standard deviation;

[0021] Calculate the absolute difference between the first mean and the second mean to obtain the mean absolute difference; calculate the absolute difference between the first standard deviation and the second standard deviation to obtain the standard deviation absolute difference;

[0022] The network similarity between the pipeline to be monitored and the corresponding monitored pipeline is obtained based on the mean absolute difference and the standard deviation absolute difference. The calculation formula is: , where is the network similarity, and are the mean absolute difference and standard deviation absolute difference, respectively. , They are and The preset scaling factor of , All are greater than 0;

[0023] The network similarity between the pipeline to be monitored and each monitored pipeline in the monitored pipeline set is calculated.

[0024] Optionally, the monitored pipeline set is screened according to the structural similarity and the network similarity, and the target monitoring pipeline set is obtained including:

[0025] Compare the structural similarity of the pipeline to be monitored and each monitored pipeline in the monitored pipeline set with a preset structural similarity threshold, and compare the network similarity with a preset network similarity threshold;

[0026] When the structural similarity is greater than a preset structural similarity threshold and the network similarity is greater than a preset network similarity threshold, the corresponding monitored pipeline in the monitored pipeline set is taken as the target monitored pipeline to obtain the target monitored pipeline set.

[0027] Optionally, obtaining the remaining power of each drone in the target drone set and the image upload quality characteristics when monitoring the corresponding monitored pipeline to obtain the priority index of the corresponding drone includes:

[0028] Obtain each image uploaded by each drone in the target drone set during the process of monitoring the corresponding monitored pipeline, and decompose it into different frequency sub-bands through wavelet transform using Daubechies wavelet function;

[0029] For each frequency subband, calculate the grayscale histogram , represents the grayscale histogram, represents a frequency subband obtained after wavelet decomposition, Indicates gray level;

[0030] Calculate the information entropy of each subband, the calculation formula is:

[0031] ;

[0032] In the formula, Indicates The information entropy of the subbands is Represents each gray level Contribution to subband entropy;

[0033] Calculate the information entropy index of the entire image. The calculation formula is: , where is the information entropy index of the entire image, is the total number of subbands;

[0034] Calculate the mean information entropy index of all images uploaded by each drone during the monitoring of the corresponding monitored pipeline as the image information entropy coefficient;

[0035] The priority selection index of the corresponding drone is obtained according to the remaining power of each drone and the image information entropy coefficient of the uploaded image when monitoring the corresponding monitored pipeline.

[0036] Optionally, obtaining the priority index of the corresponding drone according to the remaining power of each drone and the image information entropy coefficient of the uploaded image when monitoring the corresponding monitored pipeline includes:

[0037] Grayscale each image uploaded by each drone in the target drone set during the process of monitoring the corresponding monitored pipeline, and convert the color image into a grayscale image;

[0038] Divide the grayscale image into Small regions (such as 3×3 or 5×5 pixel blocks), each of which is used to calculate the local blur;

[0039] For each small area, the standard deviation of the pixel values ​​in the area is calculated, and the fuzzy index of the corresponding image is calculated based on the standard deviation of the pixel values ​​in each small area. The calculation formula is: , where is the fuzzy index, Indicates The standard deviation of the pixel values ​​in a small area;

[0040] Calculate the average fuzzy index of all images uploaded by each drone during the monitoring of the corresponding monitored pipeline as the image fuzzy coefficient;

[0041] The priority index of the corresponding drone is obtained according to the remaining power of each drone and the image information entropy coefficient and image fuzzy coefficient of the uploaded image when monitoring the corresponding monitored pipeline.

[0042] Optionally, the priority index of the corresponding drone is obtained according to the remaining power of each drone and the image information entropy coefficient and image fuzziness coefficient of the uploaded image when monitoring the corresponding monitored pipeline, including:

[0043] Obtain the remaining power of each drone and the power consumption rate during the monitoring of the corresponding monitored pipeline, and divide the remaining power by the power consumption rate to obtain the remaining usable time of the corresponding drone;

[0044] The remaining available time, image information entropy coefficient, and image fuzziness coefficient are normalized, and the priority index of the corresponding drone is obtained according to the normalized image information entropy coefficient and image fuzziness coefficient. The calculation formula is:

[0045] ;

[0046] In the formula, To prioritize the index, , , are the remaining available time after normalization, image information entropy coefficient and image fuzzy coefficient, respectively. , , They are , , The preset scaling factor of , , Both are greater than 0.

[0047] Beneficial effects of the present invention:

[0048] The present invention proposes a pipeline monitoring and maintenance method based on drones, which obtains the structural data of the pipeline to be monitored and the surrounding network signal strength data, calculates the structural similarity and network similarity between the pipeline to be monitored and each pipeline in the monitored pipeline set, and selects the target monitoring pipeline set and the corresponding drone set with similar structure and network conditions. Then, the priority index of each target drone is calculated by the remaining power and image upload quality characteristics; finally, the drone with the highest priority index is selected to perform the monitoring task, and whether the pipeline needs maintenance is determined based on the collected data; in this way, for the pipeline that introduces drones for pipeline monitoring for the first time, it is possible to select a suitable drone to monitor the pipeline to be monitored according to the actual situation, so that the monitoring data is more comprehensive and accurate, and the fine cracks or leakage points of the pipeline can be found in time, and the pipeline can be maintained in time, reducing the possibility of pipeline risks. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] The present invention will be further described below in conjunction with the accompanying drawings.

[0050] Figure 1 The figure is a flow chart of a pipeline monitoring and maintenance method based on drones. DETAILED DESCRIPTION

[0051] The following will be combined with the 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 without creative work are within the scope of protection of the present invention.

[0052] Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative work shall fall within the scope of protection of the present invention.

[0053] The embodiment of the present invention provides a pipeline monitoring and maintenance method based on a drone. Figure 1 , Figure 1 A flowchart of a pipeline monitoring and maintenance method based on a drone is provided in an embodiment of the present invention. The method comprises the following steps:

[0054] Obtain the self-structure data of the pipeline to be monitored, and record all pipelines that have been monitored by the drone as a monitored pipeline set, and calculate the structural similarity between the pipeline to be monitored and each monitored pipeline according to the self-structure data of the pipeline to be monitored and the self-structure data of each monitored pipeline in the monitored pipeline set;

[0055] Acquire network signal strength data around the pipeline to be monitored, and calculate the network similarity between the pipeline to be monitored and each monitored pipeline according to the network signal strength data of the pipeline to be monitored and the network signal strength data of each monitored pipeline in the monitored pipeline set;

[0056] The monitored pipeline set is screened according to the structural similarity and network similarity to obtain the target monitoring pipeline set, and the drone corresponding to each monitoring pipeline in the target monitoring pipeline set is taken as the target drone to obtain the target drone set;

[0057] Obtain the remaining power of each drone in the target drone set and the image upload quality characteristics when monitoring the corresponding monitored pipeline to obtain the priority selection index of the corresponding drone;

[0058] The drone with the largest index in the target drone set is selected as the final drone to perform flight monitoring on the pipeline to be monitored, and whether the pipeline to be monitored needs maintenance is determined based on the monitoring data.

[0059] Based on a pipeline monitoring and maintenance method based on a drone provided in an embodiment of the present invention, through the above method, for a pipeline that introduces a drone for pipeline monitoring for the first time, a suitable drone can be selected according to the actual situation to monitor the pipeline to be monitored, so that the monitoring data is more comprehensive and has higher accuracy, and slight cracks or leakage points in the pipeline can be discovered in time, and the pipeline can be maintained in time, reducing the possibility of risks in the pipeline.

[0060] In one embodiment, calculating the structural similarity of the pipeline to be monitored and each monitored pipeline in the monitored pipeline set according to the self-structure data of the pipeline to be monitored and the self-structure data of each monitored pipeline in the monitored pipeline set includes:

[0061] For the pipeline to be monitored and each monitored pipeline in the monitored pipeline set, preprocessing the self-structure data of the pipeline to be monitored and the self-structure data of the corresponding monitored pipeline;

[0062] Construct a distance matrix, in which each element is the distance between the self-structure data point of the pipeline to be monitored and the self-structure data point of the monitored pipeline;

[0063] The shortest path in the distance matrix is ​​calculated using a dynamic programming algorithm to find the best match between the self-structure data sequence of the pipeline to be monitored and the self-structure data sequence of the monitored pipeline;

[0064] According to the calculated shortest path, the self-structure data sequence of the pipeline to be monitored and the self-structure data sequence of the monitored pipeline are aligned;

[0065] According to the aligned self-structure data sequence of the pipeline to be monitored and the self-structure data sequence of the monitored pipeline, the similarity between them is calculated, and the calculated similarity is used as the structural similarity between the corresponding monitored pipeline and the pipeline to be monitored, so as to obtain the structural similarity of the pipeline to be monitored and each monitored pipeline in the monitored pipeline set.

[0066] It should be noted that the structural data of the pipeline to be monitored and the monitored pipeline include features such as pipeline diameter, wall thickness, material type, anti-corrosion coating, pipeline layout, etc. Specifically, similar pipeline diameter and wall thickness mean that the drone can maintain consistency in flight altitude, shooting angle and sensor sensitivity to ensure monitoring effect; Acquisition method: It can be obtained from design drawings and construction records; The same or similar materials may exhibit similar thermal conduction and reflection characteristics during monitoring, affecting the data quality of thermal imaging and optical sensors carried by drones; Acquisition method: It can be obtained by consulting construction documents; Similar anti-corrosion coatings may be relatively close in optical and thermal properties to ensure consistent reading effects of drone sensors; Acquisition method: It can be obtained by checking the construction records of anti-corrosion coatings; Similar pipeline layouts and inclinations can affect the monitoring path and the flight mode of the drone, making the monitoring results in similar environments more comparable; Acquisition method: It can be obtained through surveying and mapping records after construction; Through the similarity of these features, it can be decided to use the same drone for monitoring of two pipelines; Pipelines with similar structures can improve monitoring efficiency and data quality, reduce costs, simplify data analysis, and enhance fault detection capabilities, thereby ensuring the effectiveness and reliability of monitoring.

[0067] It should be noted that in the above steps, preprocessing: normalizing the structural data of the pipeline to be monitored and the monitored pipeline, such as standardization or normalization, to make the data comparable and reduce the impact of scale differences;

[0068] Constructing a distance matrix: For the structural data sequence of the pipeline to be monitored and the structural data sequence of each monitored pipeline, construct a distance matrix, in which each element represents the distance between the corresponding data points of the two pipelines (such as Euclidean distance or Manhattan distance);

[0069] Dynamic programming to solve the shortest path: Use the Dynamic Time Warping (DTW) algorithm or a similar dynamic programming algorithm on the distance matrix to calculate the shortest path between the two pipeline data sequences and find the best matching path between them;

[0070] Data sequence alignment: Align the structural data sequences of the two pipelines according to the shortest path, so that the differences between the two can be reasonably mapped in the sequence to ensure the rationality of the match;

[0071] Calculate similarity: After alignment, calculate the similarity between the two sequences (such as by similarity index, normalized distance, etc.), and use this as the structural similarity between the pipeline to be monitored and each monitored pipeline;

[0072] In one implementation method, through such steps, the structural similarity between the pipeline to be monitored and the monitored pipeline can be accurately evaluated, providing a more targeted reference for drone selection, thereby effectively improving monitoring accuracy.

[0073] In one embodiment, calculating the network similarity between the pipeline to be monitored and each monitored pipeline according to the network signal strength data of the pipeline to be monitored and the network signal strength data of each monitored pipeline in the monitored pipeline set includes:

[0074] Continuously acquiring network signal strength data of the pipeline to be monitored within a preset time period at multiple times to obtain a network signal strength sequence of the pipeline to be monitored;

[0075] Calculate the mean and standard deviation of the network signal strength sequence of the pipeline to be monitored as the first mean and first standard deviation;

[0076] Continuously acquiring network signal strength data of each monitored pipeline in the monitored pipeline set within a preset time period at multiple times to obtain a network signal strength sequence of the pipeline to be monitored; calculating the mean and standard deviation of the network signal strength sequence of the monitored pipeline as a second mean and a second standard deviation;

[0077] Calculate the absolute difference between the first mean and the second mean to obtain the mean absolute difference; calculate the absolute difference between the first standard deviation and the second standard deviation to obtain the standard deviation absolute difference;

[0078] The network similarity between the pipeline to be monitored and the corresponding monitored pipeline is obtained based on the mean absolute difference and the standard deviation absolute difference. The calculation formula is: , where is the network similarity, and are the mean absolute difference and standard deviation absolute difference, respectively. , They are and The preset scaling factor of , All are greater than 0;

[0079] The network similarity between the pipeline to be monitored and each monitored pipeline in the monitored pipeline set is calculated.

[0080] It should be noted that and It is set by professionals according to the actual situation. Generally, and The sum of is 1, for example and They can be 0.4, 0.6, or other numbers respectively, and are not specifically limited.

[0081] It should be noted that the preset time period is set by professionals based on actual conditions, and no specific restrictions or elaborations are made; in addition, the network signal strength data specifically refers to the strength and quality indicators of wireless signals (such as Wi-Fi, LTE, 5G, etc.) around the pipeline. These data can reflect the communication capabilities and connection stability with drones within the pipeline monitoring area; the network signal strength data of the pipeline to be monitored and the monitored pipeline can be obtained by deploying wireless signal detectors around the pipeline to be monitored and the monitored pipeline. These devices can record the surrounding signal strength in real time, including signals in different frequency bands, or other acquisition methods, which are not limited or elaborated on.

[0082] It should be noted that when the network similarity between the pipeline to be monitored and each monitored pipeline in the monitored pipeline set is large, it means that the solution of using the same drone to monitor the two pipelines is more feasible. The reason is that when the network similarity between the pipeline to be monitored and each monitored pipeline in the monitored pipeline set is large, it means that they are similar in network signal strength performance, indicating that under similar environments and conditions, the monitoring requirements of these pipelines and the operating requirements of drones are also consistent. Using the same drone for monitoring can ensure that the successful experience and data obtained by the drone on the monitored pipeline can be effectively applied to the pipeline to be monitored, thereby improving monitoring efficiency and accuracy. In addition, the similarity of network signal strength also helps drones maintain stable connections and data transmission during operation, reduce potential signal interference risks, and further enhance the reliability of monitoring.

[0083] In one embodiment, the monitored pipeline set is screened according to the structural similarity and the network similarity, and the target monitoring pipeline set is obtained, including:

[0084] Compare the structural similarity of the pipeline to be monitored and each monitored pipeline in the monitored pipeline set with a preset structural similarity threshold, and compare the network similarity with a preset network similarity threshold;

[0085] When the structural similarity is greater than a preset structural similarity threshold and the network similarity is greater than a preset network similarity threshold, the corresponding monitored pipeline in the monitored pipeline set is taken as the target monitored pipeline to obtain the target monitored pipeline set.

[0086] It should be noted that the preset structure similarity threshold and the preset network similarity threshold are set by professionals according to actual conditions, and are not specifically limited or elaborated;

[0087] It should be noted that by comparing the structural similarity and network similarity of the pipeline to be monitored with the monitored pipeline set with the preset thresholds, the target monitoring pipelines with similar structures and environmental conditions can be effectively screened out; and the drone corresponding to each monitoring pipeline in the target monitoring pipeline set is used as the target drone to obtain the target drone set. The advantage of this is that it can ensure that the monitoring capability of the target drone matches the characteristics of the pipeline to be monitored, thereby improving the accuracy and efficiency of monitoring; when the structure and network environment of the target monitoring pipeline are similar to those of the monitored pipeline, previous data and experience can be effectively applied to reduce debugging and adaptation time; in addition, matching drones working under similar conditions can provide more reliable monitoring results and reduce the risk of failures and false alarms, thereby optimizing resource allocation and monitoring costs while ensuring pipeline safety.

[0088] In one embodiment, obtaining the remaining power of each drone in the target drone set and the image upload quality characteristics when monitoring the corresponding monitored pipeline to obtain the priority index of the corresponding drone includes:

[0089] Obtain each image uploaded by each drone in the target drone set during the process of monitoring the corresponding monitored pipeline, and decompose it into different frequency sub-bands through wavelet transform using Daubechies wavelet function;

[0090] For each frequency subband, calculate the grayscale histogram , represents the grayscale histogram, represents a frequency subband obtained after wavelet decomposition, Indicates gray level;

[0091] Calculate the information entropy of each subband, the calculation formula is:

[0092] ;

[0093] In the formula, Indicates The information entropy of the subbands is Represents each gray level Contribution to subband entropy;

[0094] Calculate the information entropy index of the entire image. The calculation formula is: , where is the information entropy index of the entire image, is the total number of subbands;

[0095] Calculate the mean information entropy index of all images uploaded by each drone during the monitoring of the corresponding monitored pipeline as the image information entropy coefficient;

[0096] The priority selection index of the corresponding drone is obtained according to the remaining power of each drone and the image information entropy coefficient of the uploaded image when monitoring the corresponding monitored pipeline.

[0097] It should be noted that the image information entropy coefficient represents the complexity of the image information uploaded by each drone during the monitoring of the corresponding monitored pipeline, also known as the image information volume or image complexity, which reflects the richness of the details and information contained in the image. The larger the information entropy index, the better the quality of the uploaded image, because the larger the information entropy index, the richer the details and information contained in the image, reflecting higher image clarity and contrast; such high-quality images can more effectively capture the slight changes, defects or potential problems of the pipeline, helping drones to more accurately identify and analyze the pipeline status during the monitoring process. Therefore, by calculating the information entropy coefficient of each drone's uploaded image, its performance in the monitoring task can be objectively evaluated, and then a more suitable drone can be selected to ensure the reliability and effectiveness of the monitoring data and improve the efficiency of the entire monitoring process.

[0098] It should be noted that the images uploaded by each drone during the monitoring of the corresponding monitored pipeline can be analyzed by using image processing tools and software (such as MATLAB, OpenCV, etc.), the grayscale histogram and information entropy of each sub-band can be calculated, and the complexity and richness of the information in the image can be quantified.

[0099] In one implementation, the reason why Daubechies wavelet function is used to perform wavelet transform decomposition into different frequency sub-bands is that Daubechies wavelet has excellent time-frequency localization characteristics, which can effectively capture edge and detail information in the image, while having less redundancy and good compression performance. By decomposing the image into multiple frequency sub-bands, the image features at different frequencies can be analyzed more clearly, thereby improving the accuracy of information entropy calculation and ensuring a comprehensive evaluation of image quality; this multi-scale analysis method enables the images uploaded by drones when uploading pipeline images to more effectively identify minor defects and changes, thereby improving the accuracy and reliability of monitoring.

[0100] In one embodiment, the priority index of the corresponding drone is obtained according to the remaining power of each drone and the image information entropy coefficient of the uploaded image when monitoring the corresponding monitored pipeline, including:

[0101] Grayscale each image uploaded by each drone in the target drone set during the process of monitoring the corresponding monitored pipeline, and convert the color image into a grayscale image;

[0102] Divide the grayscale image into Small regions (such as 3×3 or 5×5 pixel blocks), each of which is used to calculate the local blur;

[0103] For each small area, the standard deviation of the pixel values ​​in the area is calculated, and the fuzzy index of the corresponding image is calculated based on the standard deviation of the pixel values ​​in each small area. The calculation formula is: , where is the fuzzy index, Indicates The standard deviation of the pixel values ​​in a small area;

[0104] Calculate the average fuzzy index of all images uploaded by each drone during the monitoring of the corresponding monitored pipeline as the image fuzzy coefficient;

[0105] The priority index of the corresponding drone is obtained according to the remaining power of each drone and the image information entropy coefficient and image fuzzy coefficient of the uploaded image when monitoring the corresponding monitored pipeline.

[0106] It should be noted that the larger the image blur coefficient, the higher the blur of the image, that is, the worse the image quality; the principle of this coefficient is based on the standard deviation of the local area: the lower the local standard deviation in the image, the blurrier the image is in that area (because the pixel value changes less). By taking the inverse sum of the standard deviations of these areas, a higher blur coefficient means that there are more areas in the image with smaller pixel changes, resulting in a higher overall blur;

[0107] It should be noted that the image blur coefficient indicates the degree of blur of the image uploaded by each drone during the monitoring of the corresponding monitored pipeline. The greater the blur of the image, the worse the quality of the uploaded image. Because when the image blur coefficient is large, it means that there are fewer edges and details in the image, or the clarity is low, resulting in the pipeline's characteristic details not being clearly displayed. This will affect the accuracy and recognizability of drone monitoring, making it difficult to detect and analyze pipeline conditions. Clear images are essential for accurately monitoring and identifying potential problems on pipelines. Therefore, the larger the blur coefficient, the worse the image quality, which may lead to unreliable monitoring results, thus affecting monitoring efficiency and effectiveness.

[0108] It should be noted that the data involved in the calculation of the image blur coefficient can be automatically extracted through image processing algorithms, usually using grayscale conversion, image segmentation, pixel value calculation and other processing steps. Specifically, after the image is grayscaled, a local window (such as a 3×3 or 5×5 pixel block) is applied to divide the entire image, and then the standard deviation of the pixel values ​​of each small area is calculated. Image processing libraries such as OpenCV or PIL can help complete these operations, thereby extracting the pixel statistics required for the calculation of the blur coefficient.

[0109] In one embodiment, the priority index of the corresponding drone is obtained according to the remaining power of each drone and the image information entropy coefficient and image fuzziness coefficient of the uploaded image when monitoring the corresponding monitored pipeline, including:

[0110] Obtain the remaining power of each drone and the power consumption rate during the monitoring of the corresponding monitored pipeline, and divide the remaining power by the power consumption rate to obtain the remaining usable time of the corresponding drone;

[0111] The remaining available time, image information entropy coefficient, and image fuzziness coefficient are normalized, and the priority index of the corresponding drone is obtained according to the normalized image information entropy coefficient and image fuzziness coefficient. The calculation formula is:

[0112] ;

[0113] In the formula, To prioritize the index, , , are the remaining available time after normalization, image information entropy coefficient and image fuzzy coefficient, respectively. , , They are , , The preset scaling factor of , , All are greater than 0;

[0114] It should be noted that , , It is set by professionals according to the actual situation. Generally, , , The sum of is 1, for example , , They can be 0.3, 0.3, 0.4 or other numbers respectively, and are not limited to any specific number.

[0115] It should be noted that the remaining power of each drone and the power consumption rate during the monitoring of the corresponding monitored pipeline can be directly obtained through the drone's internal sensors and power management system, or other acquisition methods, which are not limited or elaborated on in detail.

[0116] In one embodiment, the drone with the largest index in the target drone set is selected as the final drone to perform flight monitoring of the pipeline to be monitored, and the monitored data is used to determine whether the pipeline to be monitored needs maintenance. Specifically, the drone with the highest index is selected as the final drone to perform the flight monitoring task of the pipeline to be monitored, and images along the pipeline are collected during the flight. After the image is acquired, the pipeline surface and surrounding areas are analyzed using an image analysis algorithm, including identifying corrosion, cracks, settlement, vegetation invasion or other potential problems; the specific steps include: first pre-processing the image, such as denoising, enhancing contrast, etc., to display details more clearly; then applying edge detection, feature matching and other technologies to identify possible defect features; then, using a machine learning model to classify the defect type and assess its severity. If significant defects or potential risks are identified in key areas, the system will mark these locations as priority maintenance areas, generate automatic maintenance requests, and send reports to pipeline managers to arrange further inspections and repairs. This ensures real-time maintenance and preventive repairs of pipelines and reduces the risk of accidents.

[0117] In one implementation method, through the above method, for pipelines that introduce drones for pipeline monitoring for the first time, it is possible to select appropriate drones to monitor the pipelines to be monitored according to actual conditions, so that the monitoring data is more comprehensive and has higher accuracy, and it is possible to promptly detect slight cracks or leaks in the pipeline, and promptly maintain the pipeline, reducing the possibility of risks in the pipeline.

[0118] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A pipeline monitoring and maintenance method based on drones, characterized in that: The following steps are involved: Obtain the self-structure data of the pipeline to be monitored, and record all pipelines that have been monitored by drones as a monitored pipeline set, and calculate the structural similarity between the pipeline to be monitored and each monitored pipeline according to the self-structure data of the pipeline to be monitored and the self-structure data of each monitored pipeline in the monitored pipeline set; the self-structure data includes pipeline diameter, wall thickness, material type, anti-corrosion coating, and pipeline layout; Acquire network signal strength data around the pipeline to be monitored, and calculate the network similarity between the pipeline to be monitored and each monitored pipeline according to the network signal strength data of the pipeline to be monitored and the network signal strength data of each monitored pipeline in the monitored pipeline set; The monitored pipeline set is screened according to the structural similarity and network similarity to obtain the target monitoring pipeline set, and the drone corresponding to each monitoring pipeline in the target monitoring pipeline set is taken as the target drone to obtain the target drone set; Obtain the remaining power of each drone in the target drone set and the image upload quality characteristics when monitoring the corresponding monitored pipeline to obtain the priority selection index of the corresponding drone; The drone with the largest index is selected from the target drone set as the final drone to perform flight monitoring on the pipeline to be monitored, and whether the pipeline to be monitored needs maintenance is determined based on the monitoring data; Obtain the remaining power of each drone in the target drone set and the image upload quality characteristics when monitoring the corresponding monitored pipeline to obtain the priority index of the corresponding drone, including: Obtain each image uploaded by each drone in the target drone set during the process of monitoring the corresponding monitored pipeline, and decompose it into different frequency sub-bands through wavelet transform using Daubechies wavelet function; For each frequency subband, calculate the grayscale histogram , represents the grayscale histogram, represents a frequency subband obtained after wavelet decomposition, Indicates gray level; Calculate the information entropy of each subband, the calculation formula is: ； In the formula, Indicates The information entropy of the subbands is Represents each gray level Contribution to subband entropy; Calculate the mean information entropy index of all images uploaded by each drone during the monitoring of the corresponding monitored pipeline as the image information entropy coefficient; The priority selection index of the corresponding drone is obtained according to the remaining power of each drone and the image information entropy coefficient of the uploaded image when monitoring the corresponding monitored pipeline.

2. The pipeline monitoring and maintenance method based on drone according to claim 1 is characterized in that: Calculating the structural similarity of the pipeline to be monitored and each monitored pipeline in the monitored pipeline set based on the self-structure data of the pipeline to be monitored and the self-structure data of each monitored pipeline in the monitored pipeline set includes: For the pipeline to be monitored and each monitored pipeline in the monitored pipeline set, preprocessing the self-structure data of the pipeline to be monitored and the self-structure data of the corresponding monitored pipeline; Construct a distance matrix, in which each element is the distance between the self-structure data point of the pipeline to be monitored and the self-structure data point of the monitored pipeline; The shortest path in the distance matrix is ​​calculated using a dynamic programming algorithm to find the best match between the self-structure data sequence of the pipeline to be monitored and the self-structure data sequence of the monitored pipeline; According to the calculated shortest path, the self-structure data sequence of the pipeline to be monitored and the self-structure data sequence of the monitored pipeline are aligned; According to the aligned self-structure data sequence of the pipeline to be monitored and the self-structure data sequence of the monitored pipeline, the similarity between them is calculated, and the calculated similarity is used as the structural similarity between the corresponding monitored pipeline and the pipeline to be monitored, so as to obtain the structural similarity of the pipeline to be monitored and each monitored pipeline in the monitored pipeline set.

3. The pipeline monitoring and maintenance method based on drone according to claim 1 is characterized in that: Calculating the network similarity between the pipeline to be monitored and each monitored pipeline according to the network signal strength data of the pipeline to be monitored and the network signal strength data of each monitored pipeline in the monitored pipeline set includes: Continuously acquiring network signal strength data of the pipeline to be monitored within a preset time period at multiple times to obtain a network signal strength sequence of the pipeline to be monitored; Calculate the mean and standard deviation of the network signal strength sequence of the pipeline to be monitored as the first mean and first standard deviation; Continuously acquiring network signal strength data of each monitored pipeline in the monitored pipeline set within a preset time period at multiple times to obtain a network signal strength sequence of the pipeline to be monitored; calculating the mean and standard deviation of the network signal strength sequence of the monitored pipeline as a second mean and a second standard deviation; Calculate the absolute difference between the first mean and the second mean to obtain the mean absolute difference; calculate the absolute difference between the first standard deviation and the second standard deviation to obtain the standard deviation absolute difference; The network similarity between the pipeline to be monitored and the corresponding monitored pipeline is obtained based on the mean absolute difference and the standard deviation absolute difference. The calculation formula is: , where is the network similarity, and are the mean absolute difference and standard deviation absolute difference, respectively. , They are and The preset scaling factor of , All are greater than 0; The network similarity between the pipeline to be monitored and each monitored pipeline in the monitored pipeline set is calculated.

4. The pipeline monitoring and maintenance method based on drone according to claim 1 is characterized in that: The monitored pipeline set is screened according to the structural similarity and network similarity, and the target monitoring pipeline set includes: Compare the structural similarity of the pipeline to be monitored and each monitored pipeline in the monitored pipeline set with a preset structural similarity threshold, and compare the network similarity with a preset network similarity threshold; When the structural similarity is greater than a preset structural similarity threshold and the network similarity is greater than a preset network similarity threshold, the corresponding monitored pipeline in the monitored pipeline set is taken as the target monitored pipeline to obtain the target monitored pipeline set.

5. The pipeline monitoring and maintenance method based on drone according to claim 1 is characterized in that: According to the remaining power of each drone and the image information entropy coefficient of the uploaded image when monitoring the corresponding monitored pipeline, the priority index of the corresponding drone is obtained, including: Grayscale each image uploaded by each drone in the target drone set during the process of monitoring the corresponding monitored pipeline, and convert the color image into a grayscale image; Divide the grayscale image into For each small area, the standard deviation of the pixel values ​​in the area is calculated, and the fuzzy index of the corresponding image is calculated according to the standard deviation of the pixel values ​​in each small area. The calculation formula is: , where is the fuzzy index, Indicates The standard deviation of the pixel values ​​in a small area; Calculate the average fuzzy index of all images uploaded by each drone during the monitoring of the corresponding monitored pipeline as the image fuzzy coefficient; The priority index of the corresponding drone is obtained according to the remaining power of each drone and the image information entropy coefficient and image fuzzy coefficient of the uploaded image when monitoring the corresponding monitored pipeline.

6. The pipeline monitoring and maintenance method based on drone according to claim 5 is characterized in that: According to the remaining power of each drone and the image information entropy coefficient and image fuzziness coefficient of the uploaded image when monitoring the corresponding monitored pipeline, the priority index of the corresponding drone is obtained, including: Obtain the remaining power of each drone and the power consumption rate during the monitoring of the corresponding monitored pipeline, and divide the remaining power by the power consumption rate to obtain the remaining usable time of the corresponding drone; The remaining available time, image information entropy coefficient, and image fuzziness coefficient are normalized, and the priority index of the corresponding drone is obtained according to the normalized image information entropy coefficient and image fuzziness coefficient. The calculation formula is: ； In the formula, To prioritize the index, , , are the remaining available time after normalization, image information entropy coefficient and image fuzzy coefficient, respectively. , , They are , , The preset scaling factor of , , Both are greater than 0.

Images