Natural gas pipeline network safety operation and maintenance system based on digitization

Abstract

The invention provides a natural gas pipeline network safety operation and maintenance system based on digitization. The natural gas pipeline network safety operation and maintenance system comprises a data acquisition module, a data analysis module and a safety operation and maintenance module. Wherein the data acquisition module is used for acquiring natural gas pipeline network field image data acquired in real time; the data analysis module is used for performing local de-atomization enhancement processing on the acquired field image data to obtain an enhanced image; and performing safety analysis processing on the enhanced image by adopting a trained pipe network anomaly identification model to obtain a natural gas pipe network safety analysis result. And the safety operation and maintenance module is used for generating a corresponding safety early warning instruction according to the safety analysis result of the natural gas pipeline network. And the reliability and the effect of carrying out safe operation and maintenance on the natural gas pipeline network based on image monitoring under the abnormal environment condition can be improved.

Description

Technical Field

[0001] This invention relates to the field of natural gas pipeline network safety operation and maintenance technology, and in particular to a digital-based natural gas pipeline network safety operation and maintenance system. Background Technology

[0002] Urban natural gas pipeline networks are a crucial component of urban infrastructure, and their safe operation is of paramount importance. Traditional natural gas pipeline safety monitoring often relies on instruments such as pressure sensors and gas leak detectors, or manual inspections. However, these methods have limitations. For example, manual inspections are inefficient and have a high rate of missed detections, while sensors can only detect changes in specific parameters, and single-dimensional data often fails to reflect the true state of the natural gas pipeline network and is prone to false alarms.

[0003] Currently, there are also some natural gas pipeline safety monitoring technologies based on image monitoring technology. By setting up cameras at key locations in the pipeline network to collect images of the entire or partial pipeline network, and then performing safety analysis of the pipeline network based on the images, a fully remote and digital safety operation and maintenance operation can be carried out.

[0004] However, when using image monitoring technology for the safe operation and maintenance of natural gas pipelines, some environmental factors similar to safety incidents can easily lead to misjudgments (for example, when white mist generated by airflow drifts over the pipeline area, it is often misjudged as a pipeline leak), which affects the reliability and effectiveness of image monitoring-based safe operation and maintenance of natural gas pipelines. Summary of the Invention

[0005] To address the aforementioned problems, this invention aims to provide a digital-based natural gas pipeline network safety operation and maintenance system.

[0006] The objective of this invention is achieved through the following technical solution: This invention proposes a digital-based natural gas pipeline network safety operation and maintenance system, comprising a data acquisition module, a data analysis module, and a safety operation and maintenance module; wherein, The data acquisition module is used to acquire real-time on-site image data of the natural gas pipeline network; The data analysis module is used to perform local defogging and enhancement processing on the collected field image data to obtain enhanced images; and to perform safety analysis processing on the enhanced images using a trained pipeline anomaly identification model to obtain the natural gas pipeline safety analysis results. The safety operation and maintenance module is used to generate corresponding safety warning instructions based on the safety analysis results of the natural gas pipeline network.

[0007] Preferably, the system also includes a data display module; The data display module is used to integrate the acquired field image data and field image data into the GIS system to visualize the safety status of the natural gas pipeline network in the region.

[0008] Preferably, the on-site image data is wirelessly connected to a camera installed at the natural gas pipeline site, wherein the camera is used to collect on-site image data in the target area of ​​the natural gas pipeline in real time and transmit the collected on-site image data to the data acquisition module.

[0009] Preferably, the data analysis module includes an enhanced processing unit and a security analysis unit; wherein, The enhancement processing unit is used to perform local dehazing and enhancement processing on the acquired on-site image data to obtain an enhanced image; The safety analysis unit is used to perform safety analysis processing on the enhanced image using a trained pipeline anomaly identification model to obtain the safety analysis results of the natural gas pipeline network.

[0010] Preferably, the enhanced processing unit includes: The acquired on-site image data is processed by framing and windowing to obtain on-site image frames. ; Based on the preset calibration template, acquire on-site image frames. Region of interest and baseline pixel size ; For the region of interest, based on the two consecutive frames of scene images and Perform frame difference estimation to obtain the changing residual layer. ,in ;in This represents the pixel points in the residual layer corresponding to the change at time t. Pixel value at; and These represent the on-site image frames respectively. and medium pixel The pixel value at that location, where ; For the region of interest, further analysis is performed based on the two consecutive frames of on-site images. and Perform optical flow estimation to obtain the position change vector of each pixel. The optical flow residual layer is obtained based on the positional changes of each pixel in the region of interest. ,in ,in This represents the pixel points in the optical flow residual layer at time t. Pixel value at that location, Indicates correspondence Moment Pixel The position change vector at that location. Indicates correspondence Each pixel at time The average position change vector at that location. This indicates finding the magnitude of a vector, where ; Based on the changing residual layer and optical flow residual layer For on-site image frames Enhancement processing is performed, and the enhancement processing function used is:

[0011] In the formula, Indicates the on-site image frame Enhanced pixels Pixel value at that location, Represents on-site image frames medium pixel Pixel value at that location, This indicates the set enhancement strength factor. Represents on-site image frames The baseline pixel size, Represents pixels The region of interest function at the location, where hour, ,otherwise , This represents the pixel points in the residual layer corresponding to the change at time t. Pixel value at that location, This indicates the set standard variation pixel value. This represents the pixel points in the optical flow residual layer at time t. Pixel value at that location, This represents the set standard optical flow pixel value. and This represents the set weighting factor, where ; This represents the truncation function, where This indicates the set upper limit value; The enhanced image is obtained based on the pixel values ​​of each pixel after enhancement processing. .

[0012] Preferably, the data analysis module also includes a calibration unit; The calibration unit is used to calibrate the image after the camera is set up and the first test image is acquired, including marking the region of interest in the image. And, based on the natural gas pipeline portion of the image, mark the reference pixel size of the image. ,in , This represents the actual length of a single pixel in the natural gas pipeline section of the image. This indicates the standard actual length corresponding to a single preset pixel.

[0013] Preferably, the security analysis unit includes: Enhance the image and region of interest marker information Composition of input set ; input set The input is fed into the trained pipeline anomaly identification model, which then outputs the natural gas pipeline safety analysis results. The pipeline anomaly identification model is built on a CNN convolutional neural network structure, which includes an input layer, a backbone feature extraction network, a feature fusion network, a target detection head, a global classification head, and an output layer. The input layer is used to input the set. The backbone feature extraction network is used to extract multi-scale feature maps from the input; the feature fusion network is used to fuse features of different scales from top to bottom; the target detection head outputs the prediction results of security anomalies on feature maps of different scales; and the global classification head outputs the final security analysis results based on the prediction results of different security anomalies.

[0014] Preferably, the security operation and maintenance module includes: Based on the results of the natural gas pipeline network safety analysis, when an abnormal state is found in the natural gas pipeline network safety analysis, a corresponding early warning instruction is generated and transmitted to the management terminal.

[0015] The beneficial effects of this invention are as follows: This invention proposes a digital-based natural gas pipeline network safety operation and maintenance system. Based on image-based safety operation and maintenance of the natural gas pipeline network, the system first performs targeted defogging enhancement processing on the obtained on-site image data of the natural gas pipeline network, which helps to improve the reliability and effectiveness of image-based safety operation and maintenance of the natural gas pipeline network under abnormal environmental conditions. Attached Figure Description

[0016] The present invention will be further described with reference to the accompanying drawings, but the embodiments in the drawings do not constitute any limitation on the present invention. For those skilled in the art, other drawings can be obtained based on the following drawings without creative effort.

[0017] Figure 1This is a framework diagram of a digital-based natural gas pipeline network safety operation and maintenance system, as shown in an embodiment of the present invention. Detailed Implementation

[0018] The present invention will be further described in conjunction with the following application scenarios.

[0019] See Figure 1 This demonstrates a digitally based natural gas pipeline network safety operation and maintenance system, including a data acquisition module, a data analysis module, and a safety operation and maintenance module; among which, The data acquisition module is used to acquire real-time on-site image data of the natural gas pipeline network; The data analysis module is used to perform local defogging and enhancement processing on the collected field image data to obtain enhanced images; and to perform safety analysis processing on the enhanced images using a trained pipeline anomaly identification model to obtain the natural gas pipeline safety analysis results. The safety operation and maintenance module is used to generate corresponding safety warning instructions based on the safety analysis results of the natural gas pipeline network.

[0020] The above embodiments of the present invention propose a digital-based natural gas pipeline network safety operation and maintenance system. Based on image-based safety operation and maintenance of the natural gas pipeline network, the system first performs targeted defogging enhancement processing on the obtained on-site image data of the natural gas pipeline network, which helps to improve the reliability and effectiveness of image-based safety operation and maintenance of the natural gas pipeline network under abnormal environmental conditions.

[0021] The system described above can be built on a local server, cloud server, or other smart terminal. By acquiring real-time images from the natural gas pipeline site and performing targeted defogging processing on the image data, it can accurately distinguish between white fog caused by the on-site environment and spray phenomena caused by natural gas pipeline leaks in the images. It also improves the feature representation level of the spray area in the case of a leak, which helps to improve the accuracy and reliability of subsequent safety anomaly identification based on the on-site images.

[0022] Preferably, the system also includes a data display module; The data display module is used to integrate the acquired field image data and field image data into the GIS system to visualize the safety status of the natural gas pipeline network in the region.

[0023] When anomalies are detected in the safety analysis results, the data display module can intuitively show the location and related image information of the natural gas pipeline network where the safety anomaly is occurring. This helps managers to intuitively manage and schedule the operation and maintenance of the safety anomaly, and helps to improve the effectiveness of the safety operation and maintenance of the natural gas pipeline network.

[0024] Preferably, the on-site image data is wirelessly connected to a camera installed at the natural gas pipeline site, wherein the camera is used to collect on-site image data in the target area of ​​the natural gas pipeline in real time and transmit the collected on-site image data to the data acquisition module.

[0025] Preferably, the on-site image data carries location stamp information corresponding to the target area of ​​the natural gas pipeline network.

[0026] By installing cameras at the natural gas pipeline site, key areas in the pipeline can be monitored in real time and continuously. Based on the obtained on-site image data, a safety analysis of the natural gas pipeline can be performed, which helps to detect safety anomalies in the first instance and improve the real-time performance of the safe operation and maintenance of the natural gas pipeline.

[0027] Once the camera is set up, its position and shooting angle are fixed. When the camera is adjusted, it needs to be initialized and calibrated accordingly.

[0028] Preferably, the data analysis module includes an enhanced processing unit and a security analysis unit; wherein, The enhancement processing unit is used to perform local dehazing and enhancement processing on the acquired on-site image data to obtain an enhanced image; The safety analysis unit is used to perform safety analysis processing on the enhanced image using a trained pipeline anomaly identification model to obtain the safety analysis results of the natural gas pipeline network.

[0029] In practical applications of image-based monitoring for natural gas pipeline safety analysis, it has been found that white fog phenomena caused by cold weather can resemble pipeline leaks in images, making it difficult to distinguish between the two during model training. This often leads to misclassification of fleeting white fog as pipeline leaks, resulting in unsatisfactory image-based safety anomaly identification. To address this, this invention proposes a targeted defogging enhancement processing technique. Based on the obtained on-site image data, this technique further enhances the display of fogging conditions in pipeline leak scenarios, thereby distinguishing between pipeline leaks and environmentally generated white fog. This improves the accuracy and reliability of subsequent image-based model training and safety anomaly identification.

[0030] Preferably, the enhanced processing unit includes: The acquired on-site image data is processed by framing and windowing to obtain on-site image frames. ; Based on the preset calibration template, acquire on-site image frames. Region of interest and baseline pixel size ; For the region of interest, based on the two consecutive frames of scene images and Perform frame difference estimation to obtain the changing residual layer. ,in ;in This represents the pixel points in the residual layer corresponding to the change at time t. Pixel value at; and These represent the on-site image frames respectively. and medium pixel The pixel value at that location, where ; For the region of interest, further analysis is performed based on the two consecutive frames of on-site images. and Perform optical flow estimation to obtain the position change vector of each pixel. The optical flow residual layer is obtained based on the positional changes of each pixel in the region of interest. ,in ,in This represents the pixel points in the optical flow residual layer at time t. Pixel value at that location, Indicates correspondence Moment Pixel The position change vector at that location. Indicates correspondence Each pixel at time The average position change vector at that location. This indicates finding the magnitude of a vector, where ; Based on the changing residual layer and optical flow residual layer For on-site image frames Enhancement processing is performed, and the enhancement processing function used is:

[0031] In the formula, Indicates the on-site image frame Enhanced pixels Pixel value at that location, Represents on-site image frames medium pixel Pixel value at that location, This indicates the set enhancement strength factor. Represents on-site image frames The baseline pixel size, Represents pixels The region of interest function at the location, where hour, ,otherwise , This represents the pixel points in the residual layer corresponding to the change at time t. Pixel value at that location, This indicates the set standard variation pixel value. This represents the pixel points in the optical flow residual layer at time t. Pixel value at that location, This represents the set standard optical flow pixel value. and This represents the set weighting factor, where ; This represents the truncation function, where This indicates the set upper limit value; The enhanced image is obtained based on the pixel values ​​of each pixel after enhancement processing. .

[0032] The present invention proposes a technical solution for targeted dehazing enhancement processing of on-site image data. First, the region of interest (ROI) and the baseline pixel size information of the image are extracted as the basis for subsequent enhancement processing. For the ROI, the pixel change residuals generated by the image frame difference are statistically analyzed to characterize the pixel change features in the current image, thereby representing the haze features in the image. Furthermore, the motion of pixels is statistically analyzed using optical flow based on the image frame difference, thereby characterizing the movement features of the hazy portion in the image. Considering the atomization phenomenon caused by pipeline leaks, and given that the atomized portion ejects outward from the leak in the pipeline, the motion characteristics of its physical portion differ from those of ordinary environmental fog (the atomized portion from a pipeline leak moves unevenly and rapidly, while ordinary environmental fog moves evenly and slowly). Therefore, based on these characteristics, the motion features of the pixels in the physical portion are extracted specifically, and targeted enhancement processing is performed based on these motion features. This improves the pixel values ​​of the atomized portion of the pipeline leak, enabling the enhanced image to differentiate between the fogged portion in the leak state and the environmental state, increasing the response intensity of its edge features, texture features, and motion features. This generates sufficient features to distinguish between the two, improving the reliability and accuracy of subsequent leak detection. The proposed enhancement processing function further incorporates a baseline pixel size to adapt to the atomization characteristics caused by pipeline leaks in real-world conditions, improving adaptability to images at different distances and resolutions.

[0033] Meanwhile, for other general security anomalies, since they do not involve pixel movement or other foreign objects passing through the screen, the usual defogging enhancement process is applied to this part, which will not cause realism or other anomalies to this part, and therefore will not affect the identification results of other security anomalies.

[0034] In one scenario, the enhancement intensity factor is set. Used to control the enhancement increase, and to adjust the conversion relationship between the adjustment ratio and the pixel value, wherein The set standard change pixel value Reasonable values ​​can be set based on experience, among which... Alternatively, the average pixel value of all pixels in the region of interest within the statistically varying residual layer can be used to determine this. Standard optical flow pixel values ​​set Reasonable values ​​can be set based on experience, among which... Alternatively, it can be determined by averaging the pixel values ​​of all pixels in the region of interest within the optical flow residual layer. .

[0035] The upper threshold set in the truncation function is used to limit the enhancement amplitude and prevent over-enhancement. It is typically set to... , This represents the maximum value within the range of pixel values.

[0036] In order to ensure the enhancement effect, the weighting factor is set. .

[0037] Preferably, the data analysis module also includes a calibration unit; The calibration unit is used to calibrate the image after the camera is set up and the first test image is acquired, including marking the region of interest in the image. And, based on the natural gas pipeline portion of the image, mark the reference pixel size of the image. ,in , This represents the actual length of a single pixel in the natural gas pipeline section of the image. This indicates the standard actual length corresponding to a single preset pixel.

[0038] The region of interest includes the area in the field image where the natural gas pipeline flanges, valves, welds, pipes, etc. are located, and the area around them.

[0039] In one scenario, the true length corresponding to a single pixel in the natural gas pipeline section of an image can be calculated by setting a reference ruler on the natural gas pipeline section during the testing phase and determining the total pixel length occupied by the reference ruler in the image. For example, a reference ruler can be pasted or placed on the natural gas pipeline, and the true length corresponding to a single pixel can be obtained by capturing the pixel length of the reference ruler in the image.

[0040] After completing the camera's initial setup, the region of interest can be identified based on the initial image transmitted back by the camera. This makes subsequent image enhancement processing and security anomaly identification more targeted and further reduces interference from environmental factors.

[0041] Preferably, the security analysis unit includes: Enhance the image and region of interest marker information Composition of input set ; input set The input is fed into the trained pipeline anomaly identification model, which then outputs the natural gas pipeline safety analysis results. The pipeline anomaly identification model is built on a CNN convolutional neural network structure, which includes an input layer, a backbone feature extraction network, a feature fusion network, a target detection head, a global classification head, and an output layer. The input layer is used to input the set. The backbone feature extraction network is used to extract multi-scale feature maps from the input; the feature fusion network is used to fuse features of different scales from top to bottom; the target detection head outputs the prediction results of security anomalies on feature maps of different scales; and the global classification head outputs the final security analysis results based on the prediction results of different security anomalies.

[0042] Preferably, the backbone feature extraction network comprises five stages, wherein the Stem stage includes a first convolutional layer, a second convolutional layer, and a max pooling layer, wherein the first convolutional layer uses a convolutional kernel of... The stride is 2, and the number of output channels is 32; the second convolutional layer uses a convolutional kernel of type 1. The step size is 1, the number of output channels is 32, and the pooling kernel used in the max pooling layer is [missing information]. The stride is 2; Stage 1 includes a third convolutional layer, a first residual block, and a second residual block, wherein the third convolutional layer uses a convolutional kernel of type 2. The step size is 1, and the number of output channels is 64; the first residual block consists of 2 The convolutional blocks consist of two convolutional blocks, each with a stride of 1 and 64 output channels; the second residual block consists of two convolutional blocks. The stage consists of two convolutional blocks, each with a stride of 1, and 64 output channels. Stage 2 includes a third residual block and a fourth residual block. The third residual block consists of two convolutional blocks. The convolutional blocks consist of two separate blocks with strides of 2 and 1, and a total of 128 output channels; the fourth residual block consists of two convolutional blocks. The system consists of two convolutional blocks, each with a stride of 1, and 128 output channels. Stage 2 outputs the first feature map C1. Stage 3 includes a fifth residual block and a sixth residual block, with the fifth residual block consisting of two convolutional blocks. The convolutional blocks consist of two convolutional blocks with strides of 2 and 1 respectively, and the number of output channels is 256; the sixth residual block consists of 2... The system consists of two convolutional blocks, each with a stride of 1, and 256 output channels. Stage 3 outputs the second feature map C2. Stage 4 includes a seventh residual block and an eighth residual block, with the seventh residual block consisting of two... The convolutional blocks consist of two blocks with strides of 2 and 1 respectively, and the number of output channels is 512; the eighth residual block consists of 2... The system consists of two convolutional blocks, each with a stride of 1, and 512 output channels; Stage 4 outputs the third feature map C3. The feature fusion network includes a dimensionality reduction convolution module, an upsampling module, a fusion module, and a smoothing convolution module. It applies 1×1 convolutions with a stride of 1 and 256 output channels to the first feature map C1, the second feature map C2, and the third feature map C3, respectively, to obtain the fourth feature map L4, the fifth feature map L5, and the sixth feature map L6. The upsampling module upsamples the sixth feature map by a factor of 2. The fusion module adds the upsampled image to the fifth feature map and merges them. The smoothing convolution module then processes the merged image... The convolution process, with a stride of 1 and 256 output channels, yields the eighth feature map P8. The upsampling module upsamples the eighth feature map P8 by a factor of 2. The fusion module then adds the upsampled P8 to the fourth feature map and fuses them. Finally, the smoothing convolution module processes the fused image. Convolution processing with a stride of 1 and 256 output channels yields the seventh feature map P7; a smoothing convolution module processes the sixth feature map... Convolution processing with a stride of 1 and 256 output channels yields the ninth feature map P9. The target detection head operates on the seventh feature map P7, the eighth feature map P8, and the ninth feature map P9, respectively. Three sets of prior boxes are set for each scale. The head outputs the bounding box coordinates of the anomalous target, the target confidence score, and the probability of the anomalous state category. Anomalous states include: suspected air leakage, cracks, corrosion, and damage. Each set of prior boxes includes a detection convolutional block and an output block. The detection convolutional block processes the seventh feature map P7, the eighth feature map P8, and the ninth feature map P9 respectively. Convolution processing with a stride of 1 results in 256 output channels; a 1×1 convolution is applied to the output block with a stride of 1, resulting in 27 output channels. The global classification head performs classification based on the obtained ninth feature map P9. The process involves convolution with a stride of 1 and 256 output channels, global pooling, and fully connected processing. Finally, Softmax is used for classification and normalization to obtain the probability of each abnormal state, which includes: suspected air leakage, cracks, corrosion, and damage. The output layer obtains the final safety analysis results based on the probability of each abnormal state. The safety analysis results include safety, suspected gas leakage, cracks, corrosion and damage, as well as the location of the abnormal state.

[0043] Optionally, the pipeline anomaly identification model can also be based on a pre-trained YOLOV5 model, where the YOLOV5 model can be a pre-trained YOLOV5 model from existing technologies.

[0044] Based on the pipeline anomaly identification model, image recognition technology can be used to identify safety anomalies in natural gas pipelines, thereby improving the accuracy and efficiency of safety analysis.

[0045] In the training phase of the pipeline anomaly identification model, the training set used includes pipeline leakage images after the above-mentioned defogging enhancement processing and general environmental fog images. By labeling the two images accordingly, targeted training data is formed.

[0046] In addition, to improve the accuracy of the model, the training and testing sets also include other safe and abnormal images and normal images based on the above dehazing enhancement processing, thus constructing complete training and testing data.

[0047] Preferably, the security operation and maintenance module includes: Based on the results of the natural gas pipeline network safety analysis, when an abnormal state is found in the natural gas pipeline network safety analysis, a corresponding early warning instruction is generated and transmitted to the management terminal.

[0048] When a safety anomaly occurs in the natural gas pipeline network, it can notify the manager immediately, or generate corresponding handling guidelines based on the current safety anomaly situation to assist the manager in further investigating, maintaining, and eliminating the safety anomaly, thereby improving the level of intelligence in the safe operation and maintenance of the natural gas pipeline network.

[0049] It should be noted that the functional units / modules in the various embodiments of the present invention can be integrated into one processing unit / module, or each unit / module can exist physically separately, or two or more units / modules can be integrated into one unit / module. The integrated unit / module described above can be implemented in hardware or in the form of software functional units / modules.

[0050] From the above description of the embodiments, those skilled in the art will clearly understand that the embodiments described herein can be implemented in hardware, software, firmware, middleware, code, or any suitable combination thereof. For hardware implementation, the processor can be implemented in one or more of the following units: Application-Specific Integrated Circuit (ASIC), Digital Signal Processor (DSP), Digital Signal Processing Device (DSPD), Programmable Logic Device (PLD), Field-Programmable Gate Array (FPGA), processor, controller, microcontroller, microprocessor, other electronic units designed to implement the functions described herein, or combinations thereof. For software implementation, some or all of the processes of the embodiments can be implemented by a computer program instructing the associated hardware. During implementation, the program can be stored in a computer-readable medium or transmitted as one or more instructions or code on a computer-readable medium. Computer-readable media include computer storage media and communication media, wherein communication media include any medium that facilitates the transmission of a computer program from one place to another. Storage media can be any available medium accessible to a computer. Computer-readable media can include, but is not limited to, RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code having the form of instructions or data structures and accessible to a computer.

[0051] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit the scope of protection of the present invention. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should be able to analyze that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the essence and scope of the technical solutions of the present invention.

Claims

1. A digitization-based natural gas pipeline network safety operation and maintenance system, characterized in that, It includes a data acquisition module, a data analysis module, and a security operation and maintenance module; among which, The data acquisition module is used to acquire real-time on-site image data of the natural gas pipeline network; The data analysis module is used to perform local defogging and enhancement processing on the collected field image data to obtain enhanced images; and to perform safety analysis processing on the enhanced images using a trained pipeline anomaly identification model to obtain the natural gas pipeline safety analysis results. The safety operation and maintenance module is used to generate corresponding safety warning instructions based on the safety analysis results of the natural gas pipeline network.

2. The digital-based natural gas pipeline network safety operation and maintenance system according to claim 1, characterized in that, It also includes a data display module; The data display module is used to integrate the acquired field image data and field image data into the GIS system to visualize the safety status of the natural gas pipeline network in the region.

3. The digital-based natural gas pipeline network safety operation and maintenance system according to claim 1, characterized in that, The on-site image data is wirelessly connected to the camera installed at the natural gas pipeline site. The camera is used to collect on-site image data in the target area of ​​the natural gas pipeline in real time and transmit the collected on-site image data to the data acquisition module.

4. A digital-based natural gas pipeline network safety operation and maintenance system according to claim 1, characterized in that, The data analysis module includes an enhanced processing unit and a security analysis unit; among which, The enhancement processing unit is used to perform local dehazing and enhancement processing on the acquired on-site image data to obtain an enhanced image; The safety analysis unit is used to perform safety analysis processing on the enhanced image using a trained pipeline anomaly identification model to obtain the safety analysis results of the natural gas pipeline network.

5. A digital-based natural gas pipeline network safety operation and maintenance system according to claim 4, characterized in that, The enhanced processing unit includes: The acquired on-site image data is processed by framing and windowing to obtain on-site image frames. ; Based on the preset calibration template, acquire on-site image frames. Region of interest and baseline pixel size ; For the region of interest, based on the two consecutive frames of scene images and Perform frame difference estimation to obtain the changing residual layer. ,in ;in This represents the pixel points in the residual layer corresponding to the change at time t. Pixel value at; and These represent the on-site image frames respectively. and medium pixel The pixel value at that location, where ; For the region of interest, further analysis is performed based on the two consecutive frames of on-site images. and Perform optical flow estimation to obtain the position change vector of each pixel. The optical flow residual layer is obtained based on the positional changes of each pixel in the region of interest. ,in ,in This represents the pixel points in the optical flow residual layer at time t. Pixel value at that location, Indicates correspondence Moment Pixel The position change vector at that location. Indicates correspondence Each pixel at time The average position change vector at that location. This indicates finding the magnitude of a vector, where ; Based on the changing residual layer and optical flow residual layer For on-site image frames Enhancement processing is performed, and the enhancement processing function used is: In the formula, Indicates the on-site image frame Enhanced pixels Pixel value at that location, Represents on-site image frames medium pixel Pixel value at that location, This indicates the set enhancement strength factor. Represents on-site image frames The baseline pixel size, Represents pixels The region of interest function at the location, where hour, ,otherwise , This represents the pixel points in the residual layer corresponding to the change at time t. Pixel value at that location, This indicates the set standard variation pixel value. This represents the pixel points in the optical flow residual layer at time t. Pixel value at that location, This represents the set standard optical flow pixel value. and This represents the set weighting factor, where ; This represents the truncation function, where This indicates the set upper limit value; The enhanced image is obtained based on the pixel values ​​of each pixel after enhancement processing. .

6. A digital-based natural gas pipeline network safety operation and maintenance system according to claim 5, characterized in that, The data analysis module also includes a calibration unit; The calibration unit is used to calibrate the image after the camera is set up and the first test image is acquired, including marking the region of interest in the image. And, based on the natural gas pipeline portion of the image, mark the reference pixel size of the image. ,in , This represents the actual length of a single pixel in the natural gas pipeline section of the image. This indicates the standard actual length corresponding to a single preset pixel.

7. A digital-based natural gas pipeline network safety operation and maintenance system according to claim 5, characterized in that, The security analysis unit includes: Enhance the image and region of interest marker information Composition of input set ; input set The input is fed into the trained pipeline anomaly identification model, which then outputs the natural gas pipeline safety analysis results. The pipeline anomaly identification model is built on a CNN convolutional neural network structure, which includes an input layer, a backbone feature extraction network, a feature fusion network, a target detection head, a global classification head, and an output layer. The input layer is used to input the set. The backbone feature extraction network is used to extract multi-scale feature maps from the input; the feature fusion network is used to fuse features of different scales from top to bottom; the target detection head outputs the prediction results of security anomalies on feature maps of different scales; and the global classification head outputs the final security analysis results based on the prediction results of different security anomalies.

8. A digital-based natural gas pipeline network safety operation and maintenance system according to claim 5, characterized in that, The security operations and maintenance module includes: Based on the results of the natural gas pipeline network safety analysis, when an abnormal state is found in the natural gas pipeline network safety analysis, a corresponding early warning instruction is generated and transmitted to the management terminal.

Images