Intelligent monitoring system for oil gas in oil tank area
By constructing an oil-fouling leakage monitoring model in the oil tank area with distributed sensors and image acquisition modules, the problem that the existing technology cannot accurately judge the type and location of oil and gas leakage in the oil tank area is solved, and the accurate judgment of oil leakage and air leakage points in the oil tank area is achieved. Improves safety operation and maintenance efficiency.
Patent Information
- Application Number
- CN202411086924.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-08
- Publication Date
- 2024-12-06
AI Technical Summary
The existing oil and gas monitoring system in the oil tank area cannot accurately and quickly determine the type of oil and gas leakage and specific oil leakage points and air leakage points, which leads to difficulties in safe operation and maintenance, and it is difficult to determine oil leakage or air leakage in time when an accident occurs. The location of the oil tank area seriously endangers the safe operation of the oil tank area.
An intelligent oil and gas monitoring system in the oil tank area was designed. Data was collected through distributed sensors, combined with visible light and infrared light image acquisition modules, and an oil-fouling leakage monitoring model was constructed and trained to generate oil leakage and air leakage alarms, and use neural networks to judge leakage. The specific location of the oil point and the air leakage point.
Accurate and rapid judgment of oil and gas leakage, generate emergency or general alarms, and determine the specific locations of oil leakage and air leakage points, improve the safety operation and maintenance efficiency of the oil tank area, and ensure the safe operation of the oil tank area.
Smart Images

Figure CN119089262A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to oil and gas monitoring, and in particular to an intelligent oil and gas monitoring system for an oil tank area. Background Art
[0002] Containers for storing crude oil or other petroleum products are called oil tanks, which are mainly used in refineries, oil fields, oil depots and other industries. An oil tank farm consists of multiple oil tanks, and each oil tank farm generally stores one type of oil.
[0003] Oil and gas monitoring in oil tank areas is an important measure to ensure the safety of oil tank areas and prevent environmental pollution and fire accidents. The main purpose of oil and gas monitoring in oil tank areas is to monitor the concentration and distribution of oil and gas (mainly volatile organic compounds VOCs) in real time in the oil tank area to prevent safety hazards such as oil and gas leakage, accumulation, fire and explosion. At the same time, by analyzing the monitoring data, potential safety issues can be discovered and dealt with in a timely manner.
[0004] However, the existing oil and gas monitoring system in the oil tank area can only monitor the concentration and distribution of oil and gas in the oil tank area, and cannot accurately determine the type of oil and gas leakage, as well as the specific oil and gas leakage points, which brings great difficulties to the safe operation and maintenance of the oil tank area. Once an oil and gas leakage accident occurs, if the specific location of the oil and gas leakage points cannot be accurately and quickly determined, it will seriously endanger the safe operation of the oil tank area. Summary of the invention
[0005] 1. Technical issues to be solved
[0006] In view of the above-mentioned shortcomings of the prior art, the present invention provides an intelligent oil and gas monitoring system for an oil tank area, which can effectively overcome the defects of the prior art in that it is unable to accurately and quickly determine the type of oil and gas leakage, as well as the specific oil leakage points and gas leakage points.
[0007] (II) Technical solution
[0008] To achieve the above objectives, the present invention is implemented through the following technical solutions:
[0009] An intelligent oil and gas monitoring system for an oil tank area includes a server, wherein the server collects monitoring data from various groups of sensors distributed in the oil tank area through a monitoring data collection module, and determines a target leakage area in the oil tank area according to the collected monitoring data using a leakage monitoring module;
[0010] The server collects visible light images and infrared light images of all target oil tanks in the target leakage area through the visible light image acquisition module and the infrared light image acquisition module respectively; the server constructs an oil leakage monitoring model through the oil leakage monitoring model construction module, and trains the oil leakage monitoring model through the oil leakage monitoring model training module; the server generates an oil leakage alarm based on the detection result of the visible light image by the trained oil leakage monitoring model through the oil leakage alarm generation module, and determines the oil leakage point using the first leakage point judgment module;
[0011] The server processes the infrared light image through the infrared light image processing module, and uses the interference filtering module to filter the interference objects in the infrared light image according to the interference object recognition result of the visible light image. The server generates an air leakage alarm according to the infrared light image after interference filtering through the air leakage alarm generation module, and uses the second leakage point judgment module to judge the leakage point.
[0012] Preferably, the monitoring data acquisition module acquires monitoring data from various groups of sensors distributed in the oil tank area, and the leakage monitoring module determines the target leakage area in the oil tank area according to the acquired monitoring data, including:
[0013] Collect monitoring data from each group of sensors distributed in the oil tank area, and determine abnormal sensor groups based on the monitoring data;
[0014] According to the installation position relationship between the sensor groups in the oil tank area, the monitoring area to which the abnormal sensor group belongs is determined, and the monitoring area is used as the target leakage area in the oil tank area;
[0015] Each set of sensors includes an oil and gas concentration sensor.
[0016] Preferably, the oil leakage monitoring model building module builds an oil leakage monitoring model, including:
[0017] Construct a PSPNet model with MobileNetV2 as the feature extraction network as an oil spill monitoring model;
[0018] Among them, the feature extraction network MobileNetV2 extracts low-level hierarchical features and high-level hierarchical features, and performs convolution and upsampling according to the height, width and number of channels of the feature image, and fuses the low-level hierarchical features and high-level hierarchical features to obtain multi-level fusion features.
[0019] Preferably, the oil leakage monitoring model training module performs model training on the oil leakage monitoring model, including:
[0020] S1, dividing the image dataset into a training set, a validation set, and a test set according to a preset ratio;
[0021] S2, setting the loss function and optimizer of the oil spill monitoring model, and inputting the training set into the oil spill monitoring model for model training;
[0022] S3, the loss value is calculated based on the loss function, and the optimizer updates the model parameters according to the loss value and network gradient information;
[0023] S4. If the loss value is less than the preset loss threshold, the model training ends, and the current oil spill monitoring model is the trained oil spill monitoring model. Otherwise, return to S2 and continue to use the training set for model training.
[0024] S5. Input the validation set into the trained oil spill monitoring model, evaluate the generalization ability of the model by observing the performance on the validation set, and tune the hyperparameters and structure of the model;
[0025] S6. Input the test set into the optimized oil spill monitoring model to evaluate the model performance;
[0026] Among them, S1 divides the image dataset into a training set, a validation set, and a test set according to a preset ratio, including:
[0027] Oil leakage images at different locations on different oil tanks are collected, and the oil leakage areas in the oil leakage images are marked to form an image dataset. The image dataset is divided into a training set, a validation set, and a test set in a ratio of 6:2:2.
[0028] Preferably, the oil leakage alarm generating module generates an oil leakage alarm based on the detection result of the visible light image by the trained oil leakage monitoring model, including:
[0029] Performing corrosion and smoothing processing on the oil stain edge of the oil leak area in the visible light image to determine the first minimum circumscribed rectangle of the oil leak area;
[0030] After a preset detection interval, segment and detect the oil leakage area in the same visible light image to determine the second minimum circumscribed rectangle of the oil leakage area;
[0031] If the absolute value of the area difference between the first minimum circumscribed rectangle and the second minimum circumscribed rectangle is greater than a preset area difference threshold, an emergency oil leak alarm is generated; otherwise, a general oil leak alarm is generated.
[0032] Preferably, the first leakage point determination module determines the oil leakage point, including:
[0033] The minimum inscribed rectangle corresponding to the smallest oil leakage area in the visible light image is taken as the target judgment area, and the oil leakage point is determined based on the pixel value gradient of the target judgment area.
[0034] Preferably, the infrared light image processing module processes the infrared light image, including:
[0035] The original image pixel value of each pixel in the scene is used as the judgment basis, and the grayscale mean of the temperature calibration point image facing the black body is used as the interval judgment point. Each pixel is judged for the interval, and the non-uniformity correction coefficient of the corresponding interval is called to perform non-uniform noise adaptive correction on the infrared image.
[0036] Establish a transmission model of infrared radiation in the atmosphere, use the transmission model to remove the influence of the atmosphere on infrared imaging, and obtain the potential gas area with lower pixel value in the infrared image;
[0037] A pixel value gradient weight term is added to the cost function of the guided filter to make it sensitive to the change of pixel value gradient so as to enhance the edge of the infrared image.
[0038] Preferably, the interference filtering module performs interference filtering on the infrared light image according to the interference recognition result of the visible light image, including:
[0039] Downsample the visible light image and interpolate the infrared light image to make the images have the same size;
[0040] Using the trained target detection model to identify the interferers in the processed visible light image, and obtaining the first target area where the interferers are located;
[0041] By calculating the position transformation relationship between the image coordinates of the same scene point in the processed visible light image and the infrared light image, a mapping relationship between the processed visible light image and the infrared light image is obtained;
[0042] Based on the mapping relationship, in combination with the first target area where the interference object is located in the processed visible light image, the second target area where the interference object is located in the processed infrared light image is calculated, and the second target area is filtered.
[0043] Preferably, the air leakage alarm generating module generates an air leakage alarm according to the infrared light image after interference filtering, including:
[0044] Construct a background model consisting of pixel values at the same position in n consecutive infrared light images;
[0045] If a pixel in the current infrared image is judged as background, the pixel value of the corresponding position and its 8-neighborhood in the background model will be replaced with the pixel value in the current infrared image with probability p;
[0046] If the pixel difference between the pixel value of a pixel in the current infrared image and the pixel value of the corresponding position in the background model is greater than the preset pixel difference threshold for more times than the preset number threshold, the pixel in the current infrared image is considered to be leaking gas.
[0047] Preferably, the second leakage point determination module determines the leakage point, including:
[0048] The leaking gas contour is determined based on all pixels belonging to the leaking gas in the current infrared light image, and the leaking gas contour and the target oil tank contour of the target oil tank closest to the leaking gas are combined to determine the leaking point.
[0049] (III) Beneficial effects
[0050] Compared with the prior art, the intelligent oil and gas monitoring system for oil tank areas provided by the present invention has the following beneficial effects:
[0051] 1) The oil leakage monitoring model construction module constructs the oil leakage monitoring model, the oil leakage monitoring model training module trains the oil leakage monitoring model, the oil leakage alarm generation module generates an oil leakage alarm based on the detection results of the visible light image by the trained oil leakage monitoring model, and the first leakage point judgment module judges the oil leakage point, so that the neural network model can be used to accurately and quickly judge whether an oil leakage accident occurs and generate a corresponding oil leakage alarm. At the same time, according to the detection results of the neural network model, the minimum inscribed rectangle corresponding to the smallest oil leakage area in the visible light image is used as the target judgment area, and the specific location of the oil leakage point can be determined in combination with the pixel value gradient of the target judgment area;
[0052] 2) The infrared image processing module processes the infrared image, the interference filtering module filters the interference of the infrared image according to the interference recognition result of the visible light image, the gas leakage alarm generation module generates a gas leakage alarm according to the infrared image after the interference is filtered, and the second leakage point judgment module judges the leakage point. By performing image processing and interference filtering on the infrared image and replacing and comparing pixel values by constructing a background model, it can accurately and quickly determine whether a gas leakage accident occurs and generate a corresponding gas leakage alarm. At the same time, combined with the leakage gas contour in the infrared image and the target oil tank contour closest to the leaking gas, the specific location of the gas leakage point can be determined, which brings convenience to the safe operation and maintenance of the oil tank area and effectively ensures the safe operation of the oil tank area. BRIEF DESCRIPTION OF THE DRAWINGS
[0053] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0054] Figure 1 It is a system schematic diagram of the present invention;
[0055] Figure 2 It is a schematic diagram of the process of determining the oil leakage point in the present invention;
[0056] Figure 3 It is a schematic diagram of the process of determining the leakage point in the present invention. DETAILED DESCRIPTION
[0057] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0058] An intelligent oil and gas monitoring system for oil tank areas, such as Figure 1 and Figure 2 As shown, it includes a server, which collects monitoring data from each group of sensors distributed in the oil tank area through a monitoring data collection module, and uses a leakage monitoring module to determine a target leakage area in the oil tank area according to the collected monitoring data;
[0059] The server collects visible light images and infrared light images of all target oil tanks in the target leakage area through the visible light image acquisition module and the infrared light image acquisition module respectively. The server constructs an oil leakage monitoring model through the oil leakage monitoring model construction module, and trains the oil leakage monitoring model through the oil leakage monitoring model training module. The server generates an oil leakage alarm based on the detection results of the visible light image by the trained oil leakage monitoring model through the oil leakage alarm generation module, and determines the oil leakage point using the first leakage point judgment module.
[0060] ① The monitoring data acquisition module collects the monitoring data of each group of sensors distributed in the oil tank area. The leakage monitoring module determines the target leakage area in the oil tank area based on the collected monitoring data, including:
[0061] Collect monitoring data from each group of sensors distributed in the oil tank area, and determine abnormal sensor groups based on the monitoring data;
[0062] According to the installation position relationship between the sensor groups in the oil tank area, the monitoring area to which the abnormal sensor group belongs is determined, and the monitoring area is used as the target leakage area in the oil tank area;
[0063] Each set of sensors includes an oil and gas concentration sensor.
[0064] ② Oil leakage monitoring model construction module Construct an oil leakage monitoring model, including:
[0065] Construct a PSPNet model with MobileNetV2 as the feature extraction network as an oil spill monitoring model;
[0066] Among them, the feature extraction network MobileNetV2 extracts low-level hierarchical features and high-level hierarchical features, and performs convolution and upsampling according to the height, width and number of channels of the feature image, and fuses the low-level hierarchical features and high-level hierarchical features to obtain multi-level fusion features.
[0067] ③ The oil spill monitoring model training module trains the oil spill monitoring model, including:
[0068] S1, dividing the image dataset into a training set, a validation set, and a test set according to a preset ratio;
[0069] S2, setting the loss function and optimizer of the oil spill monitoring model, and inputting the training set into the oil spill monitoring model for model training;
[0070] S3, the loss value is calculated based on the loss function, and the optimizer updates the model parameters according to the loss value and network gradient information;
[0071] S4. If the loss value is less than the preset loss threshold, the model training ends, and the current oil spill monitoring model is the trained oil spill monitoring model. Otherwise, return to S2 and continue to use the training set for model training.
[0072] S5. Input the validation set into the trained oil spill monitoring model, evaluate the generalization ability of the model by observing the performance on the validation set, and tune the hyperparameters and structure of the model;
[0073] S6. Input the test set into the optimized oil spill monitoring model to evaluate the model performance;
[0074] Among them, S1 divides the image dataset into a training set, a validation set, and a test set according to a preset ratio, including:
[0075] Oil leakage images at different locations on different oil tanks are collected, and the oil leakage areas in the oil leakage images are marked to form an image dataset. The image dataset is divided into a training set, a validation set, and a test set in a ratio of 6:2:2.
[0076] ④ The oil leak alarm generation module generates an oil leak alarm based on the detection results of the visible light image based on the trained oil leak monitoring model, including:
[0077] Performing corrosion and smoothing processing on the oil stain edge of the oil leak area in the visible light image to determine the first minimum circumscribed rectangle of the oil leak area;
[0078] After a preset detection interval, segment and detect the oil leakage area in the same visible light image to determine the second minimum circumscribed rectangle of the oil leakage area;
[0079] If the absolute value of the area difference between the first minimum circumscribed rectangle and the second minimum circumscribed rectangle is greater than a preset area difference threshold, an emergency oil leak alarm is generated; otherwise, a general oil leak alarm is generated.
[0080] ⑤The first leakage point judgment module judges the oil leakage point, including:
[0081] The minimum inscribed rectangle corresponding to the smallest oil leakage area in the visible light image is taken as the target judgment area, and the oil leakage point is determined based on the pixel value gradient of the target judgment area.
[0082] In the above technical scheme, the oil leakage monitoring model construction module constructs the oil leakage monitoring model, the oil leakage monitoring model training module performs model training on the oil leakage monitoring model, the oil leakage alarm generation module generates an oil leakage alarm based on the detection result of the visible light image by the trained oil leakage monitoring model, and the first leakage point judgment module judges the oil leakage point, so that the neural network model can be used to accurately and quickly judge whether an oil leakage accident occurs and generate a corresponding oil leakage alarm. At the same time, according to the detection result of the neural network model, the minimum inscribed rectangle corresponding to the smallest oil leakage area in the visible light image is used as the target judgment area, and the specific location of the oil leakage point can be determined in combination with the pixel value gradient of the target judgment area.
[0083] like Figure 1 and Figure 3 As shown, the server processes the infrared light image through the infrared light image processing module, and uses the interference filtering module to filter the interference objects in the infrared light image according to the interference object recognition result of the visible light image. The server generates an air leakage alarm according to the infrared light image after interference filtering through the air leakage alarm generation module, and uses the second leakage point judgment module to judge the leakage point.
[0084] ① The infrared image processing module processes the infrared image, including:
[0085] The original image pixel value of each pixel in the scene is used as the judgment basis, and the grayscale mean of the temperature calibration point image facing the black body is used as the interval judgment point. Each pixel is judged for the interval, and the non-uniformity correction coefficient of the corresponding interval is called to perform non-uniform noise adaptive correction on the infrared image.
[0086] Establish a transmission model of infrared radiation in the atmosphere, use the transmission model to remove the influence of the atmosphere on infrared imaging, and obtain the potential gas area with lower pixel value in the infrared image;
[0087] A pixel value gradient weight term is added to the cost function of the guided filter to make it sensitive to the change of pixel value gradient so as to enhance the edge of the infrared image.
[0088] ② The interference filtering module performs interference filtering on the infrared light image according to the interference recognition result of the visible light image, including:
[0089] Downsample the visible light image and interpolate the infrared light image to make the images have the same size;
[0090] Using the trained target detection model to identify the interferers in the processed visible light image, and obtaining the first target area where the interferers are located;
[0091] By calculating the position transformation relationship between the image coordinates of the same scene point in the processed visible light image and the infrared light image, a mapping relationship between the processed visible light image and the infrared light image is obtained;
[0092] Based on the mapping relationship, in combination with the first target area where the interference object is located in the processed visible light image, the second target area where the interference object is located in the processed infrared light image is calculated, and the second target area is filtered.
[0093] ③ The air leakage alarm generation module generates an air leakage alarm based on the infrared light image after interference filtering, including:
[0094] Construct a background model consisting of pixel values at the same position in n consecutive infrared light images;
[0095] If a pixel in the current infrared image is judged as background, the pixel value of the corresponding position and its 8-neighborhood in the background model will be replaced with the pixel value in the current infrared image with probability p;
[0096] If the pixel difference between the pixel value of a pixel in the current infrared image and the pixel value of the corresponding position in the background model is greater than the preset pixel difference threshold for more times than the preset number threshold, the pixel in the current infrared image is considered to be leaking gas.
[0097] ④The second leakage point judgment module judges the leakage point, including:
[0098] The leaking gas contour is determined based on all pixels belonging to the leaking gas in the current infrared light image, and the leaking gas contour and the target oil tank contour of the target oil tank closest to the leaking gas are combined to determine the leaking point.
[0099] In the above technical scheme, the infrared image processing module processes the infrared light image, the interference filtering module filters the interference of the infrared light image according to the interference recognition result of the visible light image, the gas leakage alarm generation module generates a gas leakage alarm according to the infrared light image after the interference is filtered, and the second leakage point judgment module judges the leakage point. By performing image processing and interference filtering on the infrared light image and replacing and comparing pixel values by constructing a background model, it can accurately and quickly determine whether a gas leakage accident occurs and generate a corresponding gas leakage alarm. At the same time, combined with the leakage gas contour in the infrared light image and the target oil tank contour closest to the leaking gas, the specific location of the gas leakage point can be determined, thereby bringing convenience to the safe operation and maintenance of the oil tank area and effectively ensuring the safe operation of the oil tank area.
[0100] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. An intelligent oil and gas monitoring system for an oil tank area, characterized in that: The server includes a server, which collects monitoring data of each group of sensors distributed in the oil tank area through a monitoring data collection module, and uses a leakage monitoring module to determine a target leakage area in the oil tank area according to the collected monitoring data; The server collects visible light images and infrared light images of all target oil tanks in the target leakage area through the visible light image acquisition module and the infrared light image acquisition module respectively; the server constructs an oil leakage monitoring model through the oil leakage monitoring model construction module, and trains the oil leakage monitoring model through the oil leakage monitoring model training module; the server generates an oil leakage alarm based on the detection result of the visible light image by the trained oil leakage monitoring model through the oil leakage alarm generation module, and determines the oil leakage point using the first leakage point judgment module; The server processes the infrared light image through the infrared light image processing module, and uses the interference filtering module to filter the interference objects in the infrared light image according to the interference object recognition result of the visible light image. The server generates an air leakage alarm according to the infrared light image after interference filtering through the air leakage alarm generation module, and uses the second leakage point judgment module to judge the leakage point.
2. The intelligent oil and gas monitoring system for oil tank area according to claim 1 is characterized by: The monitoring data acquisition module collects monitoring data from each group of sensors distributed in the oil tank area, and the leakage monitoring module determines the target leakage area in the oil tank area according to the collected monitoring data, including: Collect monitoring data from each group of sensors distributed in the oil tank area, and determine abnormal sensor groups based on the monitoring data; According to the installation position relationship between the sensor groups in the oil tank area, the monitoring area to which the abnormal sensor group belongs is determined, and the monitoring area is used as the target leakage area in the oil tank area; Each set of sensors includes an oil and gas concentration sensor.
3. The intelligent oil and gas monitoring system for oil tank area according to claim 1 is characterized by: The oil leakage monitoring model building module builds an oil leakage monitoring model, including: Construct a PSPNet model with MobileNetV2 as the feature extraction network as an oil spill monitoring model; Among them, the feature extraction network MobileNetV2 extracts low-level hierarchical features and high-level hierarchical features, and performs convolution and upsampling according to the height, width and number of channels of the feature image, and fuses the low-level hierarchical features and high-level hierarchical features to obtain multi-level fusion features.
4. The intelligent oil and gas monitoring system for oil tank areas according to claim 3 is characterized by: The oil leakage monitoring model training module performs model training on the oil leakage monitoring model, including: S1, dividing the image dataset into a training set, a validation set, and a test set according to a preset ratio; S2, setting the loss function and optimizer of the oil spill monitoring model, and inputting the training set into the oil spill monitoring model for model training; S3, the loss value is calculated based on the loss function, and the optimizer updates the model parameters according to the loss value and network gradient information; S4. If the loss value is less than the preset loss threshold, the model training ends, and the current oil spill monitoring model is the trained oil spill monitoring model. Otherwise, return to S2 and continue to use the training set for model training. S5. Input the validation set into the trained oil spill monitoring model, evaluate the generalization ability of the model by observing the performance on the validation set, and tune the hyperparameters and structure of the model; S6. Input the test set into the optimized oil spill monitoring model to evaluate the model performance; Among them, S1 divides the image dataset into a training set, a validation set, and a test set according to a preset ratio, including: Oil leakage images at different locations on different oil tanks are collected, and the oil leakage areas in the oil leakage images are marked to form an image dataset. The image dataset is divided into a training set, a validation set, and a test set in a ratio of 6:2:
2.
5. The intelligent oil and gas monitoring system for oil tank area according to claim 4 is characterized by: The oil leakage alarm generation module generates an oil leakage alarm based on the detection result of the visible light image by the trained oil leakage monitoring model, including: Performing corrosion and smoothing processing on the oil stain edge of the oil leak area in the visible light image to determine the first minimum circumscribed rectangle of the oil leak area; After a preset detection interval, segment and detect the oil leakage area in the same visible light image to determine the second minimum circumscribed rectangle of the oil leakage area; If the absolute value of the area difference between the first minimum circumscribed rectangle and the second minimum circumscribed rectangle is greater than a preset area difference threshold, an emergency oil leak alarm is generated; otherwise, a general oil leak alarm is generated.
6. The intelligent oil and gas monitoring system for oil tank areas according to claim 5 is characterized by: The first leakage point determination module determines the oil leakage point, including: The minimum inscribed rectangle corresponding to the smallest oil leakage area in the visible light image is taken as the target judgment area, and the oil leakage point is determined based on the pixel value gradient of the target judgment area.
7. The intelligent oil and gas monitoring system for oil tank areas according to claim 1 is characterized by: The infrared light image processing module processes the infrared light image, including: The original image pixel value of each pixel in the scene is used as the judgment basis, and the grayscale mean of the temperature calibration point image facing the black body is used as the interval judgment point. Each pixel is judged for the interval, and the non-uniformity correction coefficient of the corresponding interval is called to perform non-uniform noise adaptive correction on the infrared image. Establish a transmission model of infrared radiation in the atmosphere, use the transmission model to remove the influence of the atmosphere on infrared imaging, and obtain the potential gas area with lower pixel value in the infrared image; A pixel value gradient weight term is added to the cost function of the guided filter to make it sensitive to the change of pixel value gradient so as to enhance the edge of the infrared image.
8. The intelligent oil and gas monitoring system for oil tank areas according to claim 7 is characterized by: The interference filtering module performs interference filtering on the infrared light image according to the interference recognition result of the visible light image, including: Downsample the visible light image and interpolate the infrared light image to make the images have the same size; Using the trained target detection model to identify the interferers in the processed visible light image, and obtaining the first target area where the interferers are located; By calculating the position transformation relationship between the image coordinates of the same scene point in the processed visible light image and the infrared light image, a mapping relationship between the processed visible light image and the infrared light image is obtained; Based on the mapping relationship, in combination with the first target area where the interference object is located in the processed visible light image, the second target area where the interference object is located in the processed infrared light image is calculated, and the second target area is filtered.
9. The intelligent oil and gas monitoring system for oil tank areas according to claim 8 is characterized by: The air leakage alarm generating module generates an air leakage alarm according to the infrared light image after interference filtering, including: Construct a background model consisting of pixel values at the same position in n consecutive infrared light images; If a pixel in the current infrared image is judged as background, the pixel value of the corresponding position and its 8-neighborhood in the background model will be replaced with the pixel value in the current infrared image with probability p; If the pixel difference between the pixel value of a pixel in the current infrared image and the pixel value of the corresponding position in the background model is greater than the preset pixel difference threshold for more times than the preset number threshold, the pixel in the current infrared image is considered to be leaking gas.
10. The intelligent oil and gas monitoring system for oil tank areas according to claim 9 is characterized in that: The second leakage point determination module determines the leakage point, including: The leaking gas contour is determined based on all pixels belonging to the leaking gas in the current infrared light image, and the leaking gas contour and the target oil tank contour of the target oil tank closest to the leaking gas are combined to determine the leaking point.
