Offshore oil and gas pipeline leak monitoring and dynamic picture enhanced rotating patrol system
By acquiring real-time infrared images and training convolutional neural network models, combined with dynamic image enhancement and rotational positioning, the problem of accurate assessment and maintenance plan allocation for offshore oil and gas pipeline leaks has been solved, achieving accurate leak judgment and severity assessment.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- JILIN UNIVERSITY
- Filing Date
- 2024-04-08
- Publication Date
- 2026-07-24
AI Technical Summary
Existing technologies cannot accurately assess the leak area, perform enhanced contrast analysis, or rate the severity of leaks in order to allocate maintenance plans after identifying leaks in offshore oil and gas pipelines.
The system employs an image acquisition module, an image preprocessing module, a model building module, an image data processing module, a leakage alarm module, a dynamic image enhancement module, and a rotation positioning module. Through real-time infrared image acquisition, preprocessing, convolutional neural network model training, and leakage alarm model judgment, it generates a leakage degree array and performs dynamic image enhancement and rotation positioning.
It enables accurate identification and severity assessment of leaks in offshore oil and gas pipelines, allows for the rational allocation of maintenance plans, and facilitates the identification of leak areas and their extent by staff through dynamic image enhancement and rotational positioning.
Smart Images
Figure CN118257975B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of oil and gas pipeline monitoring technology, specifically a rotating inspection system for monitoring and enhancing dynamic images of offshore oil and gas pipeline leaks. Background Technology
[0002] Offshore oil and gas pipelines are crucial channels for transporting oil and gas resources from the seabed to land, which is essential for meeting the world's growing energy demand. Compared with other modes of transportation, offshore oil and gas pipelines can reduce environmental impact, help protect marine ecosystems, and minimize negative impacts on the marine environment. Furthermore, they provide diversified energy transportation routes, helping to ensure a secure energy supply and mitigate energy security risks. However, leaks in offshore oil and gas pipelines can lead to serious safety problems such as fires, explosions, and environmental pollution. Therefore, it is necessary to monitor offshore oil and gas pipelines to promptly alert authorities and take measures in the event of leaks, thereby preventing safety accidents.
[0003] In the prior art, a natural gas pipeline leak monitoring system based on an infrared thermal imager, disclosed in publication number "CN115751203A," includes a regional thermal imaging image acquisition module, a gas leak area acquisition module, a hazard assessment module, and a leak point location module. The regional thermal imaging image acquisition module acquires a regional thermal imaging image based on a thermal imaging image and a regional image. The gas leak area acquisition module acquires a gas-labeled image based on the regional thermal imaging image using a segmentation neural network, thereby identifying the gas leak area. The hazard assessment module records the dynamic changes in the gas leak area, acquires the diffusion rate, and acquires the temperature changes during the gas leak process, assessing the hazard level of gas diffusion based on the diffusion rate and temperature changes. The leak point location module determines the location of the leak point based on the pixel value of each pixel in the overlaid image. This achieves the effect of real-time monitoring of natural gas pipelines and accurate location of leaks.
[0004] However, existing technologies still have significant shortcomings. For example, while current technologies require immediate on-site inspection after a pipeline leak is detected, the assessment only determines the probability of a leak, leading to inaccuracies. Furthermore, they lack the ability to further analyze the identified leak area using methods such as enhanced contrast or targeted photography to verify the accuracy of the assessment. Additionally, current technologies cannot classify the severity of the leak to determine the appropriate level of severity for subsequent maintenance planning. Summary of the Invention
[0005] The purpose of this invention is to provide a rotating inspection system for monitoring and enhancing dynamic images of leaks in offshore oil and gas pipelines, in order to solve the problems mentioned in the background art.
[0006] To achieve the above objectives, the present invention provides the following technical solution:
[0007] A rotating inspection system for monitoring and enhancing dynamic images of leaks in offshore oil and gas pipelines includes:
[0008] The image acquisition module includes a real-time image acquisition unit and a historical image acquisition unit. The real-time image acquisition unit is used to acquire real-time infrared image data of offshore oil and gas pipeline monitoring points at different shooting angles, and the historical image acquisition unit is used to acquire historical infrared image data of offshore oil and gas pipelines to form a historical infrared image database.
[0009] The image preprocessing module is used to perform noise removal, brightness and contrast adjustment operations on image data in the historical infrared image library and real-time infrared image data. The preprocessed historical infrared images are randomly divided into training set and test set. The unleashed areas of the historical infrared images are labeled "unleashed", the slightly leaked areas are labeled "slightly leaked", the moderately leaked areas are labeled "moderately leaked", and the heavily leaked areas are labeled "heavily leaked".
[0010] The model building module is used to build a leakage alarm model based on a convolutional neural network. The model is trained using infrared image data on the training set, and the model performance is tested using infrared image data on the test set to obtain the trained leakage alarm model.
[0011] The image data processing module is used to input preprocessed real-time infrared image data into a leakage alarm model. The leakage alarm model determines the leakage status of each area in the real-time infrared image and issues judgment signals, including "no leakage," "slight leakage," "moderate leakage," and "severe leakage." Different signals are assigned values, and a leakage degree array is generated. ;
[0012] in, , k represents the number of the real-time infrared image captured from different shooting angles, and k = 1, 2, 3, ..., K. This represents the degree of leakage in the j-th region of the oil and gas pipeline at the k-th shooting angle, where j = 1, 2, 3, ..., m;
[0013] Leakage alarm module, the leakage alarm module is used to array according to the degree of leakage Calculate the leakage alarm coefficient And based on the leakage alarm coefficient Different levels of leakage alarm signals are issued depending on the size of the leak;
[0014] The dynamic image enhancement module is used to array according to the degree of leakage in a positive correlation manner. The contrast of each region of the real-time infrared image is adjusted to generate a real-time infrared enhanced image;
[0015] Rotation positioning module, the rotation positioning module is used to convert the leakage degree array The data and preset thresholds Compare the results and obtain values in the real-time infrared image that are not less than a preset threshold. The region number is determined, and the image acquisition module is aligned with the real-time infrared image, assigning a value not less than a preset threshold. The area.
[0016] Furthermore, the image acquisition module consists of several mid-infrared cameras.
[0017] Furthermore, the model building module includes a model definition unit, a model training unit, a model testing unit, and a leakage alarm model output unit, wherein:
[0018] The model definition unit is used to define the prototype of the leakage alarm model, including constructing an input layer, adding a convolutional layer, adding a pooling layer, and adding a fully connected layer.
[0019] The model training unit is used to initialize the parameters of the prototype leakage alarm model, define the loss function, design the learning rate to generate the leakage alarm model, and train the leakage alarm model using historical infrared images in the training set.
[0020] The model testing unit is used to test the leakage alarm model using historical infrared images in the test set to evaluate the performance of the leakage alarm model. When the model performance evaluation result reaches the predetermined performance, the leakage alarm model is input into the leakage alarm model output unit. When the model performance evaluation result does not reach the predetermined performance, the learning rate is adjusted, and the leakage alarm model with the adjusted learning rate is re-input into the model training unit for training.
[0021] A leakage alarm model output unit is used to receive leakage alarm models whose performance evaluation results have reached a predetermined performance level.
[0022] Furthermore, the leakage alarm module includes a leakage alarm coefficient calculation unit, a leakage alarm coefficient analysis unit, and a graded alarm unit, wherein:
[0023] Leakage alarm coefficient calculation unit, the leakage alarm coefficient calculation unit is used to calculate the leakage degree array Calculate the leakage alarm coefficient The calculation formula is as follows:
[0024]
[0025] in, Array representing the degree of leakage The maximum value in, Array representing the degree of leakage The number of the largest values in the list. All are preset proportional coefficients. This represents the sub-leakage alarm coefficient of the real-time infrared image captured at the k-th shooting angle. Indicates the leakage alarm coefficient The weights;
[0026] The leakage alarm coefficient analysis unit is electrically connected to the leakage alarm coefficient calculation unit and is used to calculate the leakage alarm coefficient. The size of the signal determines whether it is a no-leakage signal, a slight-leakage signal, a moderate-leakage signal, or a severe-leakage signal.
[0027] The graded alarm unit is electrically connected to the leakage alarm coefficient analysis unit. It is used to issue no alarm signal when no leakage signal is received, issue a slight leakage alarm signal when a slight leakage signal is received, issue a moderate leakage alarm signal when a moderate leakage signal is received, and issue a severe leakage alarm signal when a severe leakage signal is received.
[0028] Furthermore, the rotation positioning module includes a threshold comparison unit and a rotation mechanism, wherein:
[0029] Threshold comparison unit, the threshold comparison unit is used to compare the leakage degree array The data and preset thresholds Compare the results and obtain values in the real-time infrared image that are not less than a preset threshold. The area code;
[0030] A rotating mechanism is used to receive real-time infrared images with values not less than a preset threshold. The region number is determined, and the image acquisition module is aligned with the real-time infrared image, assigning a value not less than a preset threshold. The area is defined, and the image acquisition module is installed on the output end of the rotating mechanism.
[0031] Furthermore, the "no leakage" signal is assigned a value of 0, the "slight leakage" signal is assigned a value of 1, the "moderate leakage" signal is assigned a value of 3, and the "severe leakage" signal is assigned a value of 5.
[0032] Furthermore, the standards for the different signals emitted by the leakage alarm module are as follows:
[0033] When satisfied At this time, the leakage alarm module does not issue an alarm signal;
[0034] When satisfied At that time, the leakage alarm module will issue a mild leakage alarm signal;
[0035] When satisfied At that time, the leakage alarm module issued a moderate leakage alarm signal;
[0036] When satisfied At that time, the leakage alarm module issued a severe leakage alarm signal.
[0037] Furthermore, the dynamic image enhancement module enhances contrast by using a leakage degree array. The contrast of the area represented by the data number with a value of 0 is used as the baseline. The contrast of the area represented by the data number with a value of 1 is increased to twice the baseline, the contrast of the area represented by the data number with a value of 3 is increased to four times the baseline, and the contrast of the area represented by the data number with a value of 5 is increased to six times the baseline.
[0038] Furthermore, the preset threshold The value is 1.
[0039] Compared with the prior art, the beneficial effects of the present invention are:
[0040] The rotating inspection system for monitoring and enhancing dynamic images of leaks in offshore oil and gas pipelines of the present invention assigns values to the degree of leakage in each area of the real-time infrared image through the image data processing module. This enables the leak alarm module to issue alarm signals of different levels according to the magnitude of the leak alarm coefficient, which facilitates the subsequent allocation of maintenance plans based on the severity of the rating. The dynamic image enhancement module enhances the area of leakage in the real-time infrared image to highlight the signs of leakage, and the rotating positioning module controls the image acquisition module to be aimed at the area of leakage for targeted monitoring, which allows staff to determine whether a leak has actually occurred in the area and the extent of the leak by observation. Attached Figure Description
[0041] Figure 1 This is a unit diagram of the rotating patrol system for monitoring and enhancing dynamic images of offshore oil and gas pipeline leaks in this invention.
[0042] Figure 2 This is a unit diagram of the model building module in this invention;
[0043] Figure 3 This is a unit diagram of the leakage alarm module in this invention;
[0044] Figure 4 This is a unit diagram of the rotation positioning module in this invention. Detailed Implementation
[0045] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.
[0046] It should be noted that, unless otherwise defined, the technical or scientific terms used in this invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.
[0047] Example:
[0048] Please see Figure 1-4 This invention provides an embodiment of a rotating inspection system for monitoring and enhancing dynamic images of offshore oil and gas pipeline leaks, comprising:
[0049] The image acquisition module includes a real-time image acquisition unit and a historical image acquisition unit. The real-time image acquisition unit is used to acquire real-time infrared image data of offshore oil and gas pipeline monitoring points at different shooting angles, while the historical image acquisition unit is used to acquire historical infrared image data of offshore oil and gas pipelines to form a historical infrared image database. The image acquisition module consists of several mid-infrared cameras.
[0050] The image preprocessing module is electrically connected to the image acquisition module. It is used to perform preprocessing operations such as noise removal and adjustment of image brightness and contrast on image data in the historical infrared image library and real-time infrared image data. The preprocessed historical infrared images are randomly divided into training set and test set. The unleashed areas of the historical infrared images are labeled "unleashed", the slightly leaked areas are labeled "slightly leaked", the moderately leaked areas are labeled "moderately leaked", and the heavily leaked areas are labeled "heavily leaked".
[0051] It should be noted that the classification of the degree of leakage in offshore oil and gas pipelines into no leakage, slight leakage, moderate leakage, and severe leakage can be determined by staff based on common knowledge in the industry, such as classifying it according to the leakage rate of oil and gas in the pipeline. No restrictions are imposed here.
[0052] It should be noted that the ratio of the number of samples in the training set to the number of samples in the test set can be 70:30 or 80:20.
[0053] It should be noted that noise removal operations on infrared images can employ techniques such as filters to reduce unstructured noise in the infrared images. Adjusting the brightness and contrast of infrared images can be done using histogram equalization to unify image characteristics across different scenarios. The method for labeling different regions of historical infrared images with varying degrees of leakage is as follows: convert the historical infrared image into an array, where each data point represents the temperature distribution of a corresponding region in the historical infrared image. Each data point is then labeled according to the actual leakage situation. For example, different strings can be added to the data in the array to represent "no leakage," "slight leakage," "moderate leakage," and "severe leakage," thus achieving the effect of labeling different regions of the historical infrared image.
[0054] The model building module is electrically connected to the image preprocessing module. It is used to build a leakage alarm model based on a convolutional neural network. The model is trained using infrared image data on the training set, and the model performance is tested using infrared image data on the test set to obtain the trained leakage alarm model.
[0055] The model building module includes a model definition unit, a model training unit, a model testing unit, and a leakage alarm model output unit, wherein:
[0056] The model definition unit is used to define the prototype of the leakage alarm model, including building an input layer that receives infrared images as input data, adding a convolutional layer that extracts image features using convolution operations, adding a pooling layer for downsampling to reduce the computational cost of the model, and adding a fully connected layer for mapping the output of the convolutional layer to the final output.
[0057] It should be noted that the prototype of the leakage alarm model can adopt a convolutional neural network model. The construction of the input layer, the addition of convolutional layers, pooling layers, and fully connected layers can all adopt existing technologies to achieve the purpose of defining the convolutional neural network model. There can be multiple convolutional layers, and the complexity of the leakage alarm model can be increased by stacking convolutional layers multiple times.
[0058] The model training unit is electrically connected to the model definition unit and the image preprocessing module. It is used to initialize the parameters of the prototype leakage alarm model, define the loss function, design the learning rate to generate the leakage alarm model, and train the leakage alarm model using historical infrared images in the training set.
[0059] The parameters of the prototype leakage alarm model include initialization parameters such as convolutional layer weights, fully connected layer weights, and biases. Existing technologies can be used to initialize the parameters of the prototype leakage alarm model, which will not be elaborated here.
[0060] The design formula for the loss function is as follows:
[0061]
[0062] It should be noted that, For loss function, This represents the total number of samples. For example, if the number of historical infrared images in the training set is n1, and each historical infrared image is divided into n2 regions, then the total number of samples N is n1*n2. The actual labels for the i-th region in the training set are (here, the label "no leakage" is assigned a value of 0, "slight leakage" is assigned a value of 1, "moderate leakage" is assigned a value of 3, and "severe leakage" is assigned a value of 5). This is the output of the i-th region in the training set after processing by the leakage alarm model;
[0063] It should be noted that, A value of 0 indicates that the leakage alarm model determines that no leakage has occurred in the i-th area. A value of 1 indicates that the leakage alarm model has determined that a minor leakage has occurred in the i-th area. A value of 3 indicates that the leakage alarm model has determined that a moderate leakage has occurred in the i-th area. A value of 5 indicates that the leakage alarm model has determined that a severe leakage has occurred in the i-th area;
[0064] The learning rate represents the step size for each parameter update. The learning rate can be 0.1, 0.01, 0.001, etc., and there is no restriction here.
[0065] The method for training the leakage alarm model using historical infrared images from the training set is as follows:
[0066] 1) Calculate the predicted output of each historical infrared image in the training set after processing by the leakage alarm model;
[0067] 2) Use a loss function to calculate the difference between the leak alarm model prediction and the actual label;
[0068] 3) Calculate the gradient of the loss function with respect to the model parameters;
[0069] 4) Update the parameters of the leakage prediction model using the gradient descent algorithm. The calculation formula is as follows:
[0070]
[0071] It should be noted that, This indicates the leakage of new parameters in the prediction model. This indicates that the old parameters of the leaked prediction model have been leaked. Indicates the learning rate. This represents the gradient of the loss function with respect to the model parameters;
[0072] 5) Repeat steps 1)-4) until the predetermined number of iterations is reached or the loss is sufficiently small to generate a leakage prediction model;
[0073] It should be noted that the data samples in the training set are used to train the leakage prediction model using existing technology, which will not be elaborated here;
[0074] The model testing unit is electrically connected to the model training unit. It is used to test the leakage alarm model using historical infrared images in the test set to evaluate the performance of the leakage alarm model. When the model performance evaluation result reaches the predetermined performance, the leakage alarm model is input into the leakage alarm model output unit. When the model performance evaluation result does not reach the predetermined performance, the learning rate is adjusted, and the leakage alarm model with the adjusted learning rate is re-input into the model training unit for training.
[0075] It should be noted that the predetermined performance of the model is generally set to an evaluation accuracy of over 95%. Historical infrared images from the test set are input into the leakage alarm model for testing to evaluate the performance of the leakage alarm model on unseen samples. The learning rate is then adjusted based on the model's evaluation performance, which is currently the best technology. Specifically, a learning rate decay strategy can be used, which will not be elaborated here. By repeating the process of model training, model testing, and learning rate adjustment, the leakage alarm model reaches a satisfactory level, thereby improving the evaluation accuracy of the leakage alarm model.
[0076] The leakage alarm model output unit is electrically connected to the model test unit and is used to receive leakage alarm models that have achieved the predetermined performance based on the model performance evaluation results.
[0077] The image data processing module is electrically connected to the image preprocessing module and the leakage alarm model output unit of the model building module. It inputs the preprocessed real-time infrared image data into the leakage alarm model, which then determines the leakage status of each area in the real-time infrared image and issues judgment signals. These signals include "no leakage," "slight leakage," "moderate leakage," and "severe leakage." Values are assigned to different signals, and a leakage degree array is generated. ;
[0078] It should be noted that, Let k represent the number of the real-time infrared image captured from different shooting angles, and k = 1, 2, 3, ..., K, where K represents the number of different shooting angles. , This represents the leakage level assignment in the j-th region of the oil and gas pipeline at the k-th shooting angle, where j = 1, 2, 3, ..., m, and m represents the number of regions divided by the oil and gas pipeline. ;
[0079] As one implementation, the "no leakage" signal is assigned a value of 0, the "slight leakage" signal is assigned a value of 1, the "moderate leakage" signal is assigned a value of 3, and the "severe leakage" signal is assigned a value of 5. The larger the value, the more severe the leakage in the j-th region of the oil and gas pipeline when photographed from the k-th shooting angle.
[0080] The leakage alarm module is electrically connected to the image data processing module and is used to array data according to the degree of leakage. Calculate the leakage alarm coefficient And based on the leakage alarm coefficient Different levels of leakage alarm signals are issued depending on the size of the leak;
[0081] The leakage alarm module includes a leakage alarm coefficient calculation unit, a leakage alarm coefficient analysis unit, and a hierarchical alarm unit, wherein:
[0082] The leakage alarm coefficient calculation unit is electrically connected to the image data processing module and is used to calculate the leakage coefficient based on the degree of leakage. Calculate the leakage alarm coefficient The calculation formula is as follows:
[0083]
[0084] It should be noted that, Array representing the degree of leakage The maximum value in, Array representing the degree of leakage The number of the largest values in the list, and and All of these can be obtained through data processing software;
[0085] It should be noted that, All are preset proportional coefficients, and ,and , The specific value is generally determined by those skilled in the art based on the actual situation;
[0086] As one implementation method, The optimal value range is 0.7–0.9. The optimal value range is 0.1–0.3;
[0087] It should be noted that the leakage level array The average value of the data The larger, The largest value in The larger, and The number of largest values in The larger the value, the greater the probability and severity of a leak in the oil and gas pipeline at the k-th shooting angle; the higher the sub-leakage alarm coefficient. The larger it is, the bigger it becomes;
[0088] It should be noted that, This represents the sub-leakage alarm coefficient of the real-time infrared image captured at the k-th shooting angle. Indicates the leakage alarm coefficient The weight, The specific value is generally determined by those skilled in the art based on the actual situation, and ;
[0089] As one implementation method, there are 5 shooting angles, i.e., K=5, which are shooting at an angle of -60 degrees from the monitoring point, shooting at an angle of -30 degrees from the monitoring point, shooting directly at the monitoring point, shooting at an angle of 30 degrees from the monitoring point, and shooting at an angle of 60 degrees from the monitoring point. ;
[0090] It should be noted that, by adjusting the K sub-leakage alarm coefficients... Perform integrated analysis to obtain leakage alarm coefficients. This improved the leakage alarm coefficient. The accuracy of the data and the leakage alarm coefficient The larger the value, the more serious the oil and gas pipeline leak;
[0091] The leakage alarm coefficient analysis unit is electrically connected to the leakage alarm coefficient calculation unit and is used to calculate the leakage alarm coefficient. The size of the signal determines whether it is a no-leakage signal, a slight-leakage signal, a moderate-leakage signal, or a severe-leakage signal.
[0092] As one implementation method, based on the leakage alarm coefficient The size of the signal is determined by the following criteria for emitting different signals:
[0093]
[0094] The graded alarm unit is electrically connected to the leakage alarm coefficient analysis unit. It is used to not issue an alarm signal when no leakage signal is received, issue a slight leakage alarm signal when a slight leakage signal is received, issue a moderate leakage alarm signal when a moderate leakage signal is received, and issue a severe leakage alarm signal when a severe leakage signal is received.
[0095] It should be noted that, based on the severity of the leak, the alarm signals are divided into minor, moderate, and severe leak alarm signals. A minor leak alarm signal alerts staff that a minor leak has occurred in the oil and gas pipeline, requiring the establishment of a maintenance plan for repair. A moderate leak alarm signal alerts staff that a moderate leak has occurred in the oil and gas pipeline, requiring immediate repair. A severe leak alarm signal alerts staff that a severe leak has occurred in the oil and gas pipeline, requiring immediate repair. This tiered alarm system facilitates the appropriate scheduling of oil and gas pipeline repair plans by staff, achieving a rational allocation of offshore oil and gas pipeline leak repair schedules.
[0096] The dynamic image enhancement module, electrically connected to the image data processing module, is used to array data according to the degree of leakage in a positive correlation manner. The contrast of each region of the real-time infrared image is adjusted to generate a real-time infrared enhanced image. This enhances the areas in the real-time infrared image where leakage has occurred, highlighting the signs of leakage. This allows staff to determine whether a leakage has occurred in the area and the severity of the leakage by observing the contrast-enhanced infrared image.
[0097] As one implementation method, to leak the array The contrast of the area represented by the data number with a value of 0 is used as the baseline. The contrast of the area represented by the data number with a value of 1 is increased to twice the baseline, the contrast of the area represented by the data number with a value of 3 is increased to four times the baseline, and the contrast of the area represented by the data number with a value of 5 is increased to six times the baseline.
[0098] It should be noted that different contrast enhancement methods can be used to enhance the contrast of different regions of a real-time infrared image, which is an existing technology and will not be elaborated here.
[0099] The rotation positioning module, electrically connected to the image data processing module, is used to convert the leakage degree array... The data and preset thresholds Compare the results and obtain values in the real-time infrared image that are not less than a preset threshold. The region number is determined, and the image acquisition module is aligned with the real-time infrared image, assigning a value not less than a preset threshold. This allows for targeted monitoring of areas where leaks occur in real-time infrared images, enabling staff to determine whether a leak has actually occurred by observation.
[0100] The rotation positioning module includes a threshold comparison unit and a rotation mechanism, wherein:
[0101] The threshold comparison unit, electrically connected to the image data processing module, is used to compare the leakage degree array. The data and preset thresholds Compare the results and obtain values in the real-time infrared image that are not less than a preset threshold. The area code;
[0102] As one implementation method, a preset threshold is used. The value of is 1;
[0103] The rotating mechanism, electrically connected to the threshold comparison unit, is used to receive values in the real-time infrared image that are not less than a preset threshold. The region number is determined, and the image acquisition module is aligned with the real-time infrared image, assigning a value not less than a preset threshold. This allows the image acquisition module to specifically monitor the area of the leaking oil and gas pipeline, making it easier for staff to observe the infrared images of the leaking area.
[0104] In one implementation, the mid-infrared camera base of the image acquisition module is mounted on the output end of the rotating mechanism, thereby causing the rotating mechanism to rotate the mid-infrared camera to align with a real-time infrared image containing a value not less than a preset threshold. In the region, the rotating mechanism can consist of a two-degree-of-freedom spherical motion system and a control device for controlling the operation of the spherical motion system. The control device is electrically connected to a threshold comparison unit, and the control device receives a value in the real-time infrared image that is not less than a preset threshold. The region number is determined, and the spherical motion system is controlled to align the image acquisition module with the real-time infrared image, assigning a value not less than a preset threshold. The area.
[0105] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.
[0106] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. 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 by 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.
[0107] 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; 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, depending on actual needs.
[0108] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.
Claims
1. A rotating inspection system for monitoring and enhancing dynamic images of leaks in offshore oil and gas pipelines, characterized in that, include: The image acquisition module includes a real-time image acquisition unit and a historical image acquisition unit. The real-time image acquisition unit is used to acquire real-time infrared image data of offshore oil and gas pipeline monitoring points at different shooting angles, and the historical image acquisition unit is used to acquire historical infrared image data of offshore oil and gas pipelines to form a historical infrared image database. The image preprocessing module is used to perform noise removal, brightness and contrast adjustment operations on image data in the historical infrared image library and real-time infrared image data. The preprocessed historical infrared images are randomly divided into training set and test set. The unleashed areas of the historical infrared images are labeled "unleashed", the slightly leaked areas are labeled "slightly leaked", the moderately leaked areas are labeled "moderately leaked", and the heavily leaked areas are labeled "heavily leaked". The model building module is used to build a leakage alarm model based on a convolutional neural network. The model is trained using infrared image data on the training set, and the model performance is tested using infrared image data on the test set to obtain the trained leakage alarm model. The image data processing module is used to input preprocessed real-time infrared image data into a leakage alarm model. The leakage alarm model determines the leakage status of each area in the real-time infrared image and issues judgment signals, including "no leakage," "slight leakage," "moderate leakage," and "severe leakage." Different signals are assigned values, and a leakage degree array is generated. ; in, , k represents the number of the real-time infrared image captured from different shooting angles, and k = 1, 2, 3, ..., K. This represents the degree of leakage in the j-th region of the oil and gas pipeline at the k-th shooting angle, where j = 1, 2, 3, ..., m; Leakage alarm module, the leakage alarm module is used to array according to the degree of leakage Calculate the leakage alarm coefficient And based on the leakage alarm coefficient Different levels of leakage alarm signals are issued depending on the size of the leak; The dynamic image enhancement module is used to array according to the degree of leakage in a positive correlation manner. The contrast of each region of the real-time infrared image is adjusted to generate a real-time infrared enhanced image; Rotation positioning module, the rotation positioning module is used to convert the leakage degree array The data and preset thresholds Compare the results and obtain values in the real-time infrared image that are not less than a preset threshold. The region number is determined, and the image acquisition module is aligned with the real-time infrared image, assigning a value not less than a preset threshold. The area.
2. The rotating patrol system for monitoring and enhancing dynamic images of offshore oil and gas pipeline leaks according to claim 1, characterized in that: The image acquisition module consists of several mid-infrared cameras.
3. The rotating patrol system for monitoring and enhancing dynamic images of offshore oil and gas pipeline leaks according to claim 1, characterized in that: The model building module includes a model definition unit, a model training unit, a model testing unit, and a leakage alarm model output unit, wherein: The model definition unit is used to define the prototype of the leakage alarm model, including constructing an input layer, adding a convolutional layer, adding a pooling layer, and adding a fully connected layer. The model training unit is used to initialize the parameters of the prototype leakage alarm model, define the loss function, design the learning rate to generate the leakage alarm model, and train the leakage alarm model using historical infrared images in the training set. The model testing unit is used to test the leakage alarm model using historical infrared images in the test set to evaluate the performance of the leakage alarm model. When the model performance evaluation result reaches the predetermined performance, the leakage alarm model is input into the leakage alarm model output unit. When the model performance evaluation result does not reach the predetermined performance, the learning rate is adjusted, and the leakage alarm model with the adjusted learning rate is re-input into the model training unit for training. A leakage alarm model output unit is used to receive leakage alarm models whose performance evaluation results have reached a predetermined performance level.
4. The rotating patrol system for monitoring and enhancing dynamic images of offshore oil and gas pipeline leaks according to claim 1, characterized in that: The leakage alarm module includes a leakage alarm coefficient calculation unit, a leakage alarm coefficient analysis unit, and a graded alarm unit, wherein: Leakage alarm coefficient calculation unit, the leakage alarm coefficient calculation unit is used to calculate the leakage degree array Calculate the leakage alarm coefficient The calculation formula is as follows: ； in, Array representing the degree of leakage The maximum value in, Array representing the degree of leakage The number of the largest values in the list. All are preset proportional coefficients. This represents the sub-leakage alarm coefficient of the real-time infrared image captured at the k-th shooting angle. Indicates the leakage alarm coefficient The weights; The leakage alarm coefficient analysis unit is electrically connected to the leakage alarm coefficient calculation unit and is used to calculate the leakage alarm coefficient. The size of the signal determines whether it is a no-leakage signal, a slight-leakage signal, a moderate-leakage signal, or a severe-leakage signal. The graded alarm unit is electrically connected to the leakage alarm coefficient analysis unit. It is used to issue no alarm signal when no leakage signal is received, issue a slight leakage alarm signal when a slight leakage signal is received, issue a moderate leakage alarm signal when a moderate leakage signal is received, and issue a severe leakage alarm signal when a severe leakage signal is received.
5. The rotating patrol system for monitoring and enhancing dynamic images of offshore oil and gas pipeline leaks according to claim 1, characterized in that: The rotation positioning module includes a threshold comparison unit and a rotation mechanism, wherein: Threshold comparison unit, the threshold comparison unit is used to compare the leakage degree array The data and preset thresholds Compare the results and obtain values in the real-time infrared image that are not less than a preset threshold. The area code; A rotating mechanism is used to receive real-time infrared images with values not less than a preset threshold. The region number is determined, and the image acquisition module is aligned with the real-time infrared image, assigning a value not less than a preset threshold. The area is defined, and the image acquisition module is installed on the output end of the rotating mechanism.
6. The rotating patrol system for monitoring and enhancing dynamic images of offshore oil and gas pipeline leaks according to any one of claims 1, 4, and 5, characterized in that: The "no leakage" signal is assigned a value of 0, the "slight leakage" signal is assigned a value of 1, the "moderate leakage" signal is assigned a value of 3, and the "severe leakage" signal is assigned a value of 5.
7. The rotating patrol system for monitoring and enhancing dynamic images of offshore oil and gas pipeline leaks according to claim 6, characterized in that: The standards for the different signals emitted by the leakage alarm module are as follows: When satisfied At this time, the leakage alarm module does not issue an alarm signal; When satisfied At that time, the leakage alarm module will issue a mild leakage alarm signal; When satisfied At that time, the leakage alarm module issued a moderate leakage alarm signal; When satisfied At that time, the leakage alarm module issued a severe leakage alarm signal.
8. The rotating patrol system for monitoring and enhancing dynamic images of offshore oil and gas pipeline leaks according to claim 6, characterized in that: The dynamic image enhancement module enhances contrast by using a leakage degree array. The contrast of the area represented by the data number with a value of 0 is used as the baseline. The contrast of the area represented by the data number with a value of 1 is increased to twice the baseline, the contrast of the area represented by the data number with a value of 3 is increased to four times the baseline, and the contrast of the area represented by the data number with a value of 5 is increased to six times the baseline.
9. The rotating patrol system for monitoring and enhancing dynamic images of offshore oil and gas pipeline leaks according to claim 6, characterized in that: The preset threshold The value is 1.