Buried pipe section flaw detection system and method based on unmanned aerial vehicle

Abstract

The invention discloses a buried pipe section flaw detection system and method based on an unmanned aerial vehicle, and relates to the technical field of flaw detection. The thermal imaging sensor and the visible light camera are installed on the unmanned aerial vehicle, and the thermal imaging sensor and the visible light camera collect image data of a target buried pipe section; the positioning device is arranged in the unmanned aerial vehicle; and the ground processing system is in wireless communication connection with the unmanned aerial vehicle and is used for receiving the image data of the target buried pipe section and outputting pipeline defect area data. According to the scheme, the pipeline area can be accurately segmented, various types of defects can be identified, the defect detection precision is improved, and potential problems of the pipeline can be found in time.

Description

Technical Field

[0001] This invention relates to the field of flaw detection technology, and in particular to a flaw detection system and method for buried pipe sections based on unmanned aerial vehicles (UAVs). Background Technology

[0002] In various factory production and operations, buried pipelines are widely used for transporting industrial raw materials, products, and treating wastewater, serving as critical infrastructure to ensure normal factory production. In factory environments, buried pipelines endure complex and harsh operating conditions over extended periods. On one hand, corrosive wastewater and exhaust gases emitted during industrial production continuously erode the inner walls of the pipelines as they are transported. On the other hand, frequent operation of heavy equipment, vehicle traffic, and potential ground subsidence in some areas create external pressure and impact on the buried pipelines, making them highly susceptible to cracks, damage, and leaks.

[0003] Traditional methods for inspecting buried pipe sections have revealed numerous drawbacks in factory environments. Manual ground inspections are extremely inefficient due to the complex and noisy environment of pipe distribution within factories, making it easy for workers to overlook well-hidden pipe defects. Ground-based detection methods, limited by factory equipment layout and building obstructions, struggle to comprehensively cover all buried pipe areas, and are particularly ineffective for pipes deeply buried underground or located in confined spaces. With the continuous expansion of factory production scale and increasingly stringent safety requirements, traditional inspection technologies can no longer meet the urgent needs of factories for efficient and accurate inspection of buried pipe sections. Summary of the Invention

[0004] In view of the above, the present invention aims to provide a flaw detection system and method for buried pipe sections based on unmanned aerial vehicles (UAVs) to solve the aforementioned technical problems.

[0005] The technical solution adopted in this invention is as follows:

[0006] This invention provides a drone-based flaw detection system for buried pipe sections, comprising:

[0007] Drones;

[0008] A thermal imaging sensor and a visible light camera are installed on the drone, and the thermal imaging sensor and the visible light camera acquire image data of the target buried pipe section;

[0009] Positioning device installed inside the drone;

[0010] A ground processing system wirelessly connected to the UAV receives image data of the target buried pipe section and outputs data on the pipe defect area.

[0011] Optionally, the UAV-based buried pipe section flaw detection system also includes:

[0012] The computing module installed on the UAV processes the image data acquired by the thermal imaging sensor and the visible light camera, and outputs the pipeline defect area data to the ground processing system.

[0013] Optionally, the ground processing system includes:

[0014] The data acquisition station receives image data of the target buried pipe section;

[0015] The terminal controls the flight of the drone.

[0016] This invention also provides a method for flaw detection of buried pipe sections based on unmanned aerial vehicles (UAVs), applied to the aforementioned UAV-based flaw detection system for buried pipe sections, the method comprising:

[0017] Acquire image data of the target buried pipe section, including thermal imaging images and visible light images;

[0018] The image data of the target buried pipe section is preprocessed to obtain the image to be detected; the image to be detected includes a first image to be detected, a second image to be detected, and a third image to be detected.

[0019] The probability of a defect in the first area to be detected is greater than the probability of a defect in the second area to be detected, and the probability of a defect in the second area to be detected is greater than the probability of a defect in the third area to be detected.

[0020] The image to be detected is input into the buried pipe section flaw detection model for defect identification to obtain the pipeline defect area;

[0021] The buried pipe section flaw detection model is obtained by training a preset neural network model based on the defect image dataset of the experimental pipeline.

[0022] The defect image dataset of the experimental pipeline includes: the defect image dataset of the experimental pipeline in the first region, the defect image dataset of the experimental pipeline in the second region, and the defect image dataset of the experimental pipeline in the third region.

[0023] Optionally, the image data of the target buried pipe section is preprocessed to obtain the image to be detected, including:

[0024] Corrections are applied to the thermal imaging image and the visible light image respectively to obtain the corrected thermal imaging image and the corrected visible light image;

[0025] Based on the coordinate information of the corrected visible light image, the corrected thermal imaging image and the corrected visible light image are registered to obtain a matching image;

[0026] The matching image is divided into regions based on the preset probability of the presence of defects to obtain the image to be detected.

[0027] Optionally, the image to be inspected is input into the buried pipe section flaw detection model for defect identification to obtain the pipeline defect area, including:

[0028] The buried pipe section flaw detection model performs image segmentation in the order of the first area to be detected image, the second area to be detected image, and the third area to be detected image to obtain the pipe area image and the background image; the pipe area image includes the first area pipe image, the second area pipe image, and the third area pipe image.

[0029] Defect identification is performed based on the pipeline area image to obtain the pipeline defect area, which includes crack area, damaged area, and leakage area.

[0030] Optionally, the training process of the buried pipe section flaw detection model includes:

[0031] Obtain a dataset of defect images of the experimental pipeline;

[0032] The defect image dataset of the experimental pipeline is processed by inputting it into the input layer of a preset neural network model to obtain the first output;

[0033] The first output is processed by the shrinkage path of the preset neural network model to obtain the second output;

[0034] The second output is processed through the expansion path of the preset neural network model to obtain the third output;

[0035] The output layer of the third output input preset neural network model is processed to obtain the buried pipe section flaw detection model.

[0036] Optionally, obtain a dataset of defect images from the experimental pipeline, including:

[0037] Acquire image data of the experimental pipeline, including thermal imaging images and visible light images of the experimental pipeline;

[0038] The image data of the experimental pipeline is labeled to obtain a defect image dataset of the experimental pipeline. The labeling includes normal areas, crack areas, damaged areas, and leaking areas.

[0039] The above-described solution of the present invention has at least the following beneficial effects:

[0040] The above-described solution of the present invention includes: a drone; a thermal imaging sensor and a visible light camera mounted on the drone, the thermal imaging sensor and the visible light camera acquiring image data of a target buried pipe section; a positioning device disposed inside the drone; and a ground processing system wirelessly connected to the drone, the ground processing system receiving the image data of the target buried pipe section and outputting pipeline defect area data. The solution of the present invention can accurately segment pipeline areas and identify various types of defects, improving the accuracy of defect detection and helping to promptly detect potential pipeline problems. Attached Figure Description

[0041] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described below with reference to the accompanying drawings, wherein:

[0042] Figure 1 This is a schematic diagram of a drone-based buried pipe section flaw detection system provided in an embodiment of the present invention.

[0043] Figure 2 A flowchart of a method for detecting flaws in buried pipe sections based on unmanned aerial vehicles (UAVs) provided in an embodiment of the present invention. Detailed Implementation

[0044] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0045] This invention proposes an embodiment of a drone-based flaw detection system for buried pipe sections, specifically, as follows: Figure 1 As shown, it includes:

[0046] Drones;

[0047] A thermal imaging sensor and a visible light camera are installed on the drone, and the thermal imaging sensor and the visible light camera acquire image data of the target buried pipe section;

[0048] Positioning device installed inside the drone;

[0049] A ground processing system wirelessly connected to the UAV receives image data of the target buried pipe section and outputs data on the pipe defect area.

[0050] In this embodiment, a drone with strong endurance and excellent flight stability is selected, equipped with a high-resolution thermal imaging sensor and a visible light camera. The thermal imaging sensor is used to capture thermal images of the pipeline caused by temperature anomalies (such as heat dissipation from leaks), while the visible light camera records the surrounding environment and appearance details of the pipeline. Simultaneously, a wireless positioning information receiving and transmitting device and a GPS positioning module are provided to receive satellite positioning signals and communicate with the outside world, ensuring that the drone accurately flies to the inspection area and transmits data in real time, accurately recording the geographical location information corresponding to each set of acquired images. Furthermore, based on the distribution map of the buried pipe sections, the drone's flight route is carefully planned to ensure comprehensive coverage of the inspection area.

[0051] The drone is also equipped with a repeater and a power supply module to enhance signal transmission capabilities; the power supply module provides stable power support for all components of the drone.

[0052] The system also includes an automated airport 5G base station, which provides take-off, landing, charging, and shelter for drones, and serves as a key node for data transmission.

[0053] The drone is equipped with a computing module, and the optional computing module can be an edge computing box. The edge computing box is used to process images locally in real time, quickly complete the preprocessing steps, and reduce data transmission latency. The edge computing box can run a lightweight Unet model locally to perform preliminary segmentation and defect judgment of the image, and then transmit the results to the server for in-depth analysis and verification.

[0054] The ground processing system includes data acquisition stations, servers, and terminals. Data acquisition stations are used to collect and process data before transmitting it to the server, ensuring the stability and timeliness of data transmission. The server is responsible for evaluating the performance of the Unet model and providing feedback to the operators based on the results. The operators issue adjustment commands through the terminals. The terminals are used to control the flight of the UAV and receive visual signals for alarm processing.

[0055] The Unet model consists of an input layer, a shrinking path, an expanding path, and an output layer.

[0056] The input layer is used to receive image data acquired from the UAV thermal imaging sensor and preprocessed, providing raw information for subsequent network processing. The input image size is 572×572×1, where "572×572" represents the width and height of the image and "1" represents the number of image channels.

[0057] The shrinking path connects to the input layer and consists of a series of consecutive operations. First, multiple 3×3 convolutional (3×3cov) layers perform convolution operations by sliding the kernel across the image, extracting basic features such as edges and textures. After every two sets of 3×3 convolutional layers, a max-pooling layer is applied, using a 2×2 pooling window to downsample the feature map, halving its size while preserving key features. This reduces data volume and computational cost, enabling the network to learn more abstract, high-level features. As the network depth increases, the semantic information of the features gradually becomes richer, helping to identify the overall shape of the pipeline and the approximate characteristics of potential defects.

[0058] The expansion path corresponds to and connects to the contraction path, primarily serving to restore image resolution and fuse features. This path starts with the downsampled low-resolution feature map and progressively expands the feature map size through 2×2 upconvolutions (upsampling operations). After each upconvolution, two sets of 3×3 convolutional layers are connected for further feature extraction and optimization. Simultaneously, the expansion path, through skip connections (the horizontal lines in the diagram), concatenates the feature maps of corresponding layers in the contraction path with the upsampled feature maps, fusing feature information from different levels. This includes both the rich semantic information extracted from the contraction path and the detailed information restored by the expansion path, thus enabling more accurate segmentation of the pipeline region and defect localization.

[0059] The output layer is connected to the expansion path. The output size of the output layer is 388×388×k, where "388×388" represents the width and height of the output feature map, and "k" represents the number of output channels. The value of k is usually related to the number of categories to be identified. For example, in the flaw detection of buried pipe sections, k may correspond to the normal area of ​​the pipeline, different types of defect areas (cracks, damage, leaks, etc.), and background. The output layer processes the feature map output by the expansion path through a 1×1 convolution (1×1kcov), converting it into the final segmentation result. That is, each pixel is classified into the corresponding category, generating a segmentation map corresponding to the input image. This visually displays the location and type of defects in the buried pipe section, providing an important basis for pipeline maintenance and management.

[0060] Embodiments of the present invention also provide a method for flaw detection of buried pipe sections based on unmanned aerial vehicles (UAVs), applied to the UAV-based buried pipe section flaw detection system described in the above embodiments. The method includes:

[0061] Step 11: Acquire image data of the target buried pipe section, including thermal imaging images and visible light images;

[0062] Step 12: Preprocess the image data of the target buried pipe section to obtain the image to be detected; the image to be detected includes a first image of the area to be detected, a second image of the area to be detected, and a third image of the area to be detected.

[0063] Step 13: The probability of a defect in the first area to be detected is greater than the probability of a defect in the second area to be detected, and the probability of a defect in the second area to be detected is greater than the probability of a defect in the third area to be detected.

[0064] Step 14: Input the image to be detected into the buried pipe section flaw detection model for defect identification to obtain the pipeline defect area;

[0065] The buried pipe section flaw detection model is obtained by training a preset neural network model based on the defect image dataset of the experimental pipeline.

[0066] The defect image dataset of the experimental pipeline includes: the defect image dataset of the experimental pipeline in the first region, the defect image dataset of the experimental pipeline in the second region, and the defect image dataset of the experimental pipeline in the third region.

[0067] In this embodiment, the flight route of the UAV is carefully planned based on the distribution map of the buried pipe section to ensure full coverage of the area to be inspected; during the flight, the thermal imaging sensor and the visible light camera synchronously collect image data at a preset frequency and angle; after the collected data is preliminarily processed by the repeater, it is transmitted to the subsequent equipment by the wireless positioning information receiving and sending device.

[0068] The edge computing box performs image correction and registration: it corrects thermal and visible light images separately to eliminate image distortion caused by sensor characteristics or environmental interference. Using GPS information and image feature points recorded during image acquisition, the two types of images are accurately registered so that they correspond to the same detection area, resulting in a matched image. This process performs preliminary screening and organization of the acquired images, reducing the burden on subsequent processing.

[0069] Next, the matched image is normalized: techniques such as filtering and contrast enhancement are used to improve image clarity and quality, highlighting the characteristics of the pipeline and its surrounding environment; at the same time, the image is normalized to adjust the pixel values ​​to an appropriate range, reducing data transmission delay and facilitating subsequent Unet algorithm processing.

[0070] Furthermore, the matched image is divided into regions based on the preset probability of defect presence to obtain the image to be detected. Specifically, based on prior experience, the matched image is divided into regions according to the probability of areas in the pipeline being prone to damage, and different markings are applied to regions of different probabilities, such as red regions for high probability, yellow regions for medium probability, and green regions for low probability. After division, the image to be detected is obtained. Then, the Unet algorithm is used to segment the buried pipe section flaw detection model according to the risk level, obtaining the pipeline region image and the background image; wherein the pipeline region image includes a first region pipeline image, a second region pipeline image, and a third region pipeline image; here, the first region pipeline image is a high-risk region image, the second region pipeline image is a medium-risk region image, and the third region pipeline image is a low-risk region image.

[0071] Furthermore, the pre-processed image to be inspected is input into the trained buried pipe section flaw detection model. The model segments the image, separating the pipe area from the background and identifying pipe defect areas. Based on this, it determines whether the pipe has defects and the type and location of the defects, including crack areas, broken areas, and leak areas. The edge computing box can run a lightweight buried pipe section flaw detection model locally to perform preliminary image segmentation and defect judgment, and then transmit the results to the server for in-depth analysis and verification.

[0072] The test results are presented in an intuitive way, such as marking the location and type of pipeline defects on visible light images and generating test reports for easy viewing by staff. The terminal serves as a human-computer interaction interface, displaying the test results processed by the server to the operators in a visual form, so that they can quickly understand the pipeline condition.

[0073] Simultaneously, the performance of the buried pipe section flaw detection model is evaluated using the test dataset, and indicators such as accuracy and recall are calculated. If the model performance does not meet expectations, the training data or model parameters can be adjusted and retrained to improve detection accuracy and reliability. Here, the server is responsible for evaluating the model performance and providing feedback to the operators based on the results. The operators then issue adjustment instructions through the terminal. The 5G base station at the automated airport provides take-off, landing, charging, and shelter locations for drones. As a key node for data transmission, it utilizes the high-speed transmission characteristics of the 5G network to transmit the data initially processed and analyzed by the edge computing box to the data collection station. The data collection station collects and organizes the data before transmitting it to the server, ensuring the stability and timeliness of data transmission.

[0074] In the defect identification process, the Unet model needs to be trained first. The training process for the buried pipe section flaw detection model includes:

[0075] Obtaining a defect image dataset of the experimental pipeline: From the preprocessed image data, image samples containing different states of the pipeline, such as normal areas, cracks, damage, and leaks, are manually labeled to construct training and testing datasets. The training dataset is used to train the Unet model, and the testing dataset is used to evaluate the model's performance. In this step, the server assists in managing and storing a large amount of image data, ensuring data security and integrity.

[0076] The training dataset is input into the Unet model for training. The Unet network structure mainly consists of two parts: a contraction path and an expansion path, forming an overall "U" shape. The contraction path begins with a series of consecutive convolutional layers, usually in pairs, with each pair using a 3x3 kernel. For raw images captured by UAV thermal imaging sensors and visible light cameras, the convolutional layers can capture low-level features such as edges and textures. As the network deepens, the number of kernels gradually increases, allowing the network to learn more complex and abstract features. For example, deep convolutional layers can identify unique patterns or abnormal temperature distribution features related to pipe defects. After each pair of convolutional layers, a 2x2 max-pooling layer is immediately applied. Its function is to downsample the feature map, reduce the resolution, retain important feature information, and integrate small local features into a macroscopic feature representation, which helps in the subsequent identification of the overall shape of the pipe and potential defect areas.

[0077] The expansion path corresponds to the contraction path and begins with the upsampling layer. Upsampling aims to restore the downsampled feature map in the contraction path to the original image resolution. Common methods include deconvolution or nearest neighbor interpolation. In the flaw detection of buried pipe sections, the upsampling layer gradually enlarges the low-resolution, high-semantic feature map extracted from the contraction path, preparing for fusion with the feature map of the corresponding layer of the contraction path. After the upsampling layer is the feature fusion layer, which concatenates the upsampled feature map with the feature map of the corresponding position in the contraction path. Through this fusion, the network combines the rich contextual information (high semantic features) extracted from the contraction path with the detailed information (low-level features) gradually restored by the expansion path, thereby more accurately segmenting the pipe area and identifying defects. For example, when identifying pipe cracks, the deep feature map of the contraction path provides the overall features and semantic information of the crack, while the feature map of the corresponding layer of the expansion path contains detailed image details of the crack location. After fusion, the crack shape, length, and other parameters can be determined more accurately. After feature fusion, a series of convolutional layers using 3x3 kernels further process the fused feature map, optimizing and refining the features to enable the network to more accurately output the segmentation results of the pipe region and the identification results of defects. For example, convolutional operations further highlight the boundary between the pipe and the background, clearly outlining the contours of pipe defects. The network's final layer is the output layer, typically consisting of a 1x1 convolutional layer. This convolutional layer transforms the feature map processed by the dilated path into the final segmentation result.

[0078] In the flaw detection of buried pipe sections, the output layer outputs a segmentation map of the same size as the input image. Each pixel is labeled as belonging to a normal area, a defective area (such as cracks, damage, leaks, etc.), or a background area of ​​the pipeline, achieving accurate flaw detection of the buried pipe section. For example, the segmentation map of the output layer can intuitively display the defect location, specific type, and position on the pipeline, providing clear detection information for pipeline maintenance personnel. The training dataset is input into the Unet model for training. During training, the model parameters, such as the convolution kernel size, number of layers, and learning rate, are continuously adjusted so that the model can deeply learn the image features of the pipeline under different states, establish an accurate feature recognition pattern, and train a flaw detection model for buried pipe sections.

[0079] The UAV-based buried pipe section flaw detection system and method of this invention has the advantages of rapid coverage of large-area buried pipe sections in factories by UAVs, which greatly improves detection efficiency compared with traditional detection methods. It can also detect areas that are difficult for humans to reach, achieving comprehensive detection. Furthermore, based on the Unet algorithm, the collected images are processed and analyzed to accurately segment the pipeline area and identify various types of defects, improving the accuracy of defect detection and helping to discover potential pipeline problems in a timely manner. Finally, the detection results are presented in a visual form and a report is generated, providing intuitive information for factory pipeline maintenance and management, facilitating scientific decision-making by staff, and reducing the risk of pipeline accidents.

[0080] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0081] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0082] In the embodiments provided by this invention, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0083] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0084] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0085] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, ROM, RAM, magnetic disks, or optical disks.

[0086] Furthermore, it should be noted that in the apparatus and method of the present invention, it is obvious that the components or steps can be decomposed and / or recombined. These decompositions and / or recombinations should be considered equivalent solutions of the present invention. Moreover, the steps performing the above series of processes can naturally be executed in the order described, but are not necessarily required to be executed in chronological order; some steps can be executed in parallel or independently of each other. Those skilled in the art will understand that all or any step or component of the method and apparatus of the present invention can be implemented in any computing device (including processors, storage media, etc.) or network of computing devices, in hardware, firmware, software, or a combination thereof. This is something that those skilled in the art can achieve by using their basic programming skills after reading the description of the present invention.

[0087] Therefore, the object of the present invention can also be achieved by running a program or a set of programs on any computing device. The computing device can be a known general-purpose device. Therefore, the object of the present invention can also be achieved simply by providing a program product containing program code implementing the method or apparatus. That is, such a program product also constitutes the present invention, and the storage medium storing such a program product also constitutes the present invention. Obviously, the storage medium can be any known storage medium or any storage medium developed in the future. It should also be noted that in the apparatus and method of the present invention, it is obvious that the components or steps can be decomposed and / or recombined. These decompositions and / or recombinations should be considered equivalent to the present invention. Furthermore, the steps performing the above series of processes can naturally be performed in the order described, but are not necessarily required to be performed in chronological order. Some steps can be performed in parallel or independently of each other.

[0088] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A flaw detection system for buried pipe sections based on unmanned aerial vehicles (UAVs), characterized in that, include: Drones; A thermal imaging sensor and a visible light camera are installed on the drone, and the thermal imaging sensor and the visible light camera acquire image data of the target buried pipe section; Positioning device installed inside the drone; A ground processing system wirelessly connected to the UAV receives image data of the target buried pipe section and outputs data on the pipe defect area.

2. The UAV-based buried pipe section flaw detection system according to claim 1, characterized in that, Also includes: The computing module installed on the UAV processes the image data acquired by the thermal imaging sensor and the visible light camera, and outputs the pipeline defect area data to the ground processing system.

3. The UAV-based buried pipe section flaw detection system according to claim 2, characterized in that, The ground processing system includes: The data acquisition station receives image data of the target buried pipe section; The terminal controls the flight of the drone.

4. A method for flaw detection of buried pipe sections based on unmanned aerial vehicles (UAVs), characterized in that, The method applied to the UAV-based buried pipe section flaw detection system of claim 1 includes: Acquire image data of the target buried pipe section, including thermal imaging images and visible light images; The image data of the target buried pipe section is preprocessed to obtain the image to be detected; the image to be detected includes a first image to be detected, a second image to be detected, and a third image to be detected. The probability of a defect in the first area to be detected is greater than the probability of a defect in the second area to be detected, and the probability of a defect in the second area to be detected is greater than the probability of a defect in the third area to be detected. The image to be detected is input into the buried pipe section flaw detection model for defect identification to obtain the pipeline defect area; The buried pipe section flaw detection model is obtained by training a preset neural network model based on the defect image dataset of the experimental pipeline. The defect image dataset of the experimental pipeline includes: the defect image dataset of the experimental pipeline in the first region, the defect image dataset of the experimental pipeline in the second region, and the defect image dataset of the experimental pipeline in the third region.

5. The method for detecting flaws in buried pipe sections based on unmanned aerial vehicles (UAVs) according to claim 4, characterized in that, The image data of the target buried pipe section is preprocessed to obtain the image to be detected, including: Corrections are applied to the thermal imaging image and the visible light image respectively to obtain the corrected thermal imaging image and the corrected visible light image; Based on the coordinate information of the corrected visible light image, the corrected thermal imaging image and the corrected visible light image are registered to obtain a matching image; The matching image is divided into regions based on the preset probability of the presence of defects to obtain the image to be detected.

6. The method for detecting flaws in buried pipe sections based on unmanned aerial vehicles (UAVs) according to claim 4, characterized in that, The image to be inspected is input into the buried pipe section flaw detection model for defect identification to obtain the pipeline defect area, including: The buried pipe section flaw detection model performs image segmentation in the order of the first area to be detected image, the second area to be detected image, and the third area to be detected image to obtain the pipe area image and the background image; the pipe area image includes the first area pipe image, the second area pipe image, and the third area pipe image. Defect identification is performed based on the pipeline area image to obtain the pipeline defect area, which includes crack area, damaged area, and leakage area.

7. The method for detecting flaws in buried pipe sections based on unmanned aerial vehicles (UAVs) according to claim 4, characterized in that, The training process of the buried pipe section flaw detection model includes: Obtain a dataset of defect images of the experimental pipeline; The defect image dataset of the experimental pipeline is processed by inputting it into the input layer of a preset neural network model to obtain the first output; The first output is processed by the shrinkage path of the preset neural network model to obtain the second output; The second output is processed through the expansion path of the preset neural network model to obtain the third output; The output layer of the third output input preset neural network model is processed to obtain the buried pipe section flaw detection model.

8. The method for detecting flaws in buried pipe sections based on unmanned aerial vehicles (UAVs) according to claim 7, characterized in that, Obtain a dataset of defect images from the experimental pipeline, including: Acquire image data of the experimental pipeline, including thermal imaging images and visible light images of the experimental pipeline; The image data of the experimental pipeline is labeled to obtain a defect image dataset of the experimental pipeline. The labeling includes normal areas, crack areas, damaged areas, and leaking areas.

Images