Pipeline surrounding environment automatic monitoring method and system based on machine vision

Abstract

The invention relates to the technical field of image processing, and discloses a pipeline surrounding environment automatic monitoring method and system based on machine vision, and the system comprises the steps that an acquisition unit acquires image data in a to-be-monitored range to determine a warning range; the processing unit collects to-be-analyzed data in real time and divides the to-be-analyzed data into a plurality of sub-region data; the judgment unit processes the data of each sub-region based on a Gaussian mixture model and judges whether the data change or not; when it is judged that the change occurs, the change type is identified, and a risk factor is determined; the early warning unit determines an early warning level; and when the early warning level is a second early warning level, the adjusting unit performs accumulative monitoring on the changed sub-regions, judges whether the risk factors are adjusted or not, and updates the early warning level according to the adjusted risk factors. The monitoring range is comprehensively sensed, the warning range is dynamically determined, the monitoring precision and efficiency of the surrounding environment of the pipeline are improved, and potential safety hazards caused by personal errors and environment sudden change are avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and in particular to a method and system for automatically monitoring the surrounding environment of a pipeline based on machine vision. Background Art

[0002] With the acceleration of industrialization, pipeline transportation plays a vital role in the fields of oil, natural gas, and chemical industry. The safe operation of pipelines is directly related to the stability of production, energy supply, and environmental protection. Therefore, real-time monitoring and risk assessment of pipelines and their surroundings become particularly important. Especially in areas with complex and changing external environments of pipelines, the existence of any obstacles or changes in the external environment may lead to safety hazards such as damage, corrosion, or leakage of pipelines.

[0003] Traditional manual inspections rely on on-site inspections by staff, which is not only costly and inefficient, but also easily affected by factors such as weather, terrain, and staff fatigue, resulting in inaccurate monitoring results. Although fixed sensors can monitor certain parameters of the pipeline in real time (such as temperature, pressure, etc.), their comprehensive perception of changes in the environment around the pipeline is limited. They can only monitor environmental changes at specific locations, but cannot effectively perceive dynamic changes in the environment around the pipeline, and lack adaptive adjustment capabilities. For example, when the surrounding environment changes drastically, the monitoring system cannot respond in time, resulting in potential safety hazards not being discovered in time.

[0004] Therefore, it is necessary to design an automatic monitoring method and system for the pipeline surrounding environment based on machine vision to solve the problems existing in the current technology. Summary of the invention

[0005] In view of this, the present invention proposes a method and system for automatic monitoring of the pipeline surrounding environment based on machine vision, aiming to solve the problems existing in the current automatic monitoring technology of the pipeline surrounding environment, namely, lack of dynamic adjustment capability, time-consuming and labor-intensive monitoring process, and unclear monitoring effect.

[0006] In one aspect, the present invention provides a pipeline surrounding environment automatic monitoring system based on machine vision, comprising:

[0007] A collection unit is configured to collect image data within a range to be monitored, perform edge recognition on the image data, determine pipeline information and obstacle information, and determine a warning range according to the pipeline information and obstacle information;

[0008] A processing unit is configured to collect image data of the warning range in real time as data to be analyzed, and divide the data to be analyzed into a plurality of sub-area data;

[0009] a judgment unit configured to process each of the sub-region data based on a Gaussian mixture model and judge whether the sub-region data has changed; when the judgment unit determines that the sub-region data has changed, identify the change type and determine the risk factor according to all the sub-regions that have changed;

[0010] an early warning unit, configured to compare the risk factor with a risk threshold to determine an early warning level, wherein the early warning level includes a first early warning level, a second early warning level, and a third early warning level;

[0011] The adjustment unit is configured to, when the early warning unit determines that the early warning level is the second early warning level, perform cumulative monitoring on the sub-area where the change occurs, and determine whether to adjust the risk factor based on the cumulative monitoring results, and update the early warning level based on the adjusted risk factor.

[0012] Furthermore, when the acquisition unit determines the warning range according to the pipeline information and the obstacle information, it includes:

[0013] The pipeline information includes pipeline diameter, pipeline length and pipeline quantity;

[0014] The obstacle information includes the average density of plants, the average height of plants, and the shortest distance between plants and pipelines;

[0015] ;

[0016] Where L represents the width of the warning range, Di represents the diameter of the i-th pipeline, Li represents the length of the i-th pipeline, N represents the number of pipelines, h represents the average height of the plants, and d represents the shortest distance between the plants and the pipelines. represents the average density of plants, k represents the distance attenuation coefficient, and the value range of k is [0.5, 0.9];

[0017] The warning range is formed by extending from the center of the pipeline to both sides of the pipeline, with the extension distance on each side being the width of the warning range.

[0018] Furthermore, the judging unit processes each of the sub-region data based on the Gaussian mixture model and judges whether the sub-region data changes, including:

[0019] The judgment unit collects historical image data, and selects image data having the same position as the sub-region data from the historical image data as model building data;

[0020] The judgment unit performs image segmentation on the model building data, removes the remaining images outside the pipeline, and initializes the mean, covariance matrix and mixing weight of the Gaussian mixture model with the pipeline image;

[0021] Estimate the probability that each data point in the model establishment data belongs to each Gaussian distribution based on the initialization parameters to obtain the posterior probability;

[0022] updating model parameters according to the posterior probability;

[0023] The probability that each pixel of the sub-region data belongs to the current model is calculated, and when the likelihood of the pixel in the sub-region data is lower than a likelihood threshold, the judgment unit determines that the sub-region data has changed.

[0024] Furthermore, the calculation of the probability that each pixel of the sub-region data belongs to the current model includes:

[0025] ;

[0026] Among them, P represents the likelihood of x data in the sub-region data, M represents the number of distributions in the Gaussian mixture model, represents the mixture weight of the j-th Gaussian distribution, Represents the probability density function of the j-th Gaussian distribution.

[0027] Furthermore, when the judgment unit identifies the change type and determines the risk factor according to all the sub-regions where the change occurs, it includes:

[0028] The judgment unit performs image comparison on pixels that have changed in the sub-region data to determine the change type, where the change type includes biological change and non-biological change;

[0029] ;

[0030] Where A represents the area of ​​the changing region, Az represents the image area of ​​the sub-region, E represents the shortest distance between the changing region and the pipeline, E0 represents the safe distance of the pipeline, S represents the biological changing area, R0 represents the initial risk factor, and R0 is a positive integer.

[0031] Furthermore, the early warning unit compares the risk factor with the risk threshold to determine the early warning level, including:

[0032] The early warning unit compares the risk factor with a first risk threshold and a second risk threshold set in advance, respectively, and determines the early warning level according to the comparison result, wherein the first risk threshold is less than the second risk threshold;

[0033] When the risk factor is less than or equal to a first risk threshold, the warning unit determines the warning level as the first warning level; when the risk factor is greater than the first risk threshold and less than or equal to a second risk threshold, the warning unit determines the warning level as the second warning level; when the risk factor is greater than the second risk threshold, the warning unit determines the warning level as the third warning level; the first warning level is lower than the second warning level, and the second warning level is lower than the third warning level.

[0034] Furthermore, when the adjustment unit determines whether to adjust the risk factor according to the accumulated monitoring results, it includes:

[0035] The adjustment unit obtains an area change curve of the change region according to the accumulated monitoring results, determines a maximum slope according to the area change curve, compares the maximum slope with a slope threshold, and determines whether to adjust the risk factor according to the comparison result;

[0036] When the maximum slope is greater than the slope threshold, the adjustment unit determines to adjust the risk factor;

[0037] When the maximum slope is less than or equal to a slope threshold, the adjustment unit determines not to adjust the risk factor.

[0038] Further, when the adjustment unit determines to adjust the risk factor, it includes:

[0039] The adjustment unit performs segment processing on the region area change curve;

[0040] When the area change curve has a peak and a valley point, the adjustment unit divides the area change curve into a first curve, a second curve and a third curve at the peak and the valley point positions;

[0041] When there is only one of a peak point or a valley point in the area change curve, the adjustment unit divides the area change curve into a first curve and a third curve at the peak point or the valley point;

[0042] When the area change curve is an increasing function curve or a decreasing function curve, the adjustment unit divides the area change curve into a first curve and a third curve at a middle position of the area change curve;

[0043] The adjustment unit adjusts the risk factor according to the first curve and the third curve to obtain an adjusted risk factor.

[0044] Furthermore, the adjustment unit determines an adjustment factor according to the first curve and the third curve to adjust the risk factor, and obtains the adjusted risk factor, including:

[0045] The adjusting unit obtains an average rate of change of the area of ​​the changing region according to the area change curve;

[0046] The adjustment factor is calculated by the following formula:

[0047]

[0048] in, represents the adjustment factor, T represents the number of area change rates of the changing regions in the first curve, Bt represents the area change rate of the t-th changing region in the first curve, B0 represents the average area change rate of the changing regions, Y represents the number of area change rates of the changing regions in the second curve, and By represents the area change rate of the y-th changing region in the second curve.

[0049] Compared with the prior art, the beneficial effect of the present invention is that by combining machine vision technology and Gaussian mixture model, the automatic, real-time monitoring and risk assessment of the surrounding environment of the pipeline are realized. The acquisition unit locates the pipeline information and surrounding obstacles through edge recognition, realizes the comprehensive perception of the monitoring range, and thus dynamically determines the warning range, effectively covering the environmental changes around the pipeline. The processing unit collects image data within the warning range in real time, and divides the data into sub-areas through image segmentation, which is convenient for more detailed analysis. The judgment unit uses the Gaussian mixture model to detect changes in each sub-area, automatically judges and identifies the type of change (such as biological change or non-biological change), and evaluates potential threats by calculating risk factors. The early warning unit judges the early warning level based on the comparison between the risk factor and the preset threshold, and dynamically adjusts the monitoring strategy and early warning level according to the actual situation, ensuring that it can respond quickly when changes occur and adjust the early warning measures in time. The accuracy and efficiency of monitoring the surrounding environment of the pipeline are improved, and potential safety hazards caused by human errors and environmental mutations are avoided.

[0050] On the other hand, the present application also provides a method for automatically monitoring the surrounding environment of a pipeline based on machine vision, which is applied to the above-mentioned automatic monitoring system for the surrounding environment of a pipeline based on machine vision, comprising:

[0051] Collect image data within the range to be monitored, perform edge recognition on the image data, determine pipeline information and obstacle information, and determine the warning range based on the pipeline information and obstacle information;

[0052] Collecting image data of the warning range in real time as data to be analyzed, and dividing the data to be analyzed into a plurality of sub-area data;

[0053] Processing each of the sub-region data based on a Gaussian mixture model and determining whether the sub-region data has changed; when it is determined that the sub-region data has changed, identifying the type of change and determining a risk factor based on all the sub-regions that have changed;

[0054] Comparing the risk factor with the risk threshold to determine a warning level, wherein the warning level includes a first warning level, a second warning level, and a third warning level;

[0055] When the warning unit determines that the warning level is the second warning level, the adjustment unit performs cumulative monitoring on the sub-area where the change occurs, determines whether to adjust the risk factor according to the cumulative monitoring result, and updates the warning level according to the adjusted risk factor.

[0056] It is understandable that the above-mentioned method and system for automatic monitoring of the surrounding environment of pipelines based on machine vision have the same beneficial effects, which will not be elaborated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present invention. Moreover, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:

[0058] Figure 1 A structural block diagram of a pipeline surrounding environment automatic monitoring system based on machine vision provided by an embodiment of the present invention;

[0059] Figure 2 A flow chart of a method for automatically monitoring the surrounding environment of a pipeline based on machine vision provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0060] Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments described herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art. It should be noted that, in the absence of conflict, the embodiments of the present invention and the features described in the embodiments can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0061] In some embodiments of the present application, see Figure 1 As shown, a pipeline surrounding environment automatic monitoring system based on machine vision includes:

[0062] The acquisition unit is configured to acquire image data within the range to be monitored, perform edge recognition on the image data, determine pipeline information and obstacle information, and determine a warning range according to the pipeline information and obstacle information;

[0063] The processing unit is configured to collect image data of the warning range in real time as data to be analyzed, and divide the data to be analyzed into a plurality of sub-area data;

[0064] A judgment unit is configured to process each sub-region data based on a Gaussian mixture model and judge whether the sub-region data has changed; when the judgment unit determines that the sub-region data has changed, identify the change type and determine the risk factor according to all the sub-regions that have changed;

[0065] An early warning unit is configured to compare the risk factor with the risk threshold and determine an early warning level, wherein the early warning level includes a first early warning level, a second early warning level, and a third early warning level;

[0066] The adjustment unit is configured to, when the early warning unit determines that the early warning level is the second early warning level, perform cumulative monitoring on the sub-area where the change occurs, determine whether to adjust the risk factor based on the cumulative monitoring results, and update the early warning level based on the adjusted risk factor.

[0067] Specifically, the acquisition unit collects image data within the range to be monitored. Through high-definition cameras or sensors, image information of the pipeline and its surrounding environment is obtained in real time. After edge recognition, the image data extracts the shape, position and surrounding obstacles (such as vegetation, buildings, etc.) of the pipeline. This information helps to further define the warning range of the pipeline. Through edge recognition, the pipeline and the surrounding environment are distinguished to determine which areas need to be monitored. The processing unit further analyzes the collected image data. After determining the warning range, the image data is divided into multiple sub-areas for more detailed change monitoring. Dividing the image into sub-areas helps to more efficiently locate the specific area where the change occurs, and improves the accuracy and response speed of monitoring. The judgment unit is a key part of the system. It processes each sub-area based on the Gaussian mixture model. The Gaussian mixture model analyzes the data by maximizing the likelihood and determines whether there is a change. When it is determined that the sub-area data has changed, the judgment unit identifies the type of change (such as biological change or abiotic change) and calculates the risk factor for subsequent risk assessment. The early warning unit compares the calculated risk factor with the preset risk threshold to determine the current risk level. Warning level: first warning level (lowest risk), second warning level (medium risk), third warning level (high risk). According to the value of the risk factor, the warning unit will dynamically generate warning information. When the second warning level is determined, the adjustment unit will conduct further cumulative monitoring of the changed sub-areas. The cumulative monitoring time can be set according to actual use needs. Through continuous monitoring data, the adjustment unit evaluates whether the value of the risk factor needs to be adjusted and updates the warning level accordingly. Effectively prevent the hysteresis of the warning and ensure that the system remains highly sensitive and accurate in a long-term changing environment.

[0068] It is understandable that through machine vision technology, real-time images of the environment around the pipeline are collected, and edge recognition technology is used to accurately locate pipeline information and obstacle information. Through the Gaussian mixture model, each sub-area is dynamically processed, which can comprehensively monitor every change in the surrounding environment, avoiding the limitations and monitoring blind spots in traditional monitoring methods. Through data processing and risk factor calculation, environmental changes are automatically identified and risk assessments are performed without manual intervention. The efficiency and accuracy of pipeline monitoring are improved. The adjustment unit dynamically adjusts the risk factor according to the cumulative monitoring results, and updates the warning level according to the adjusted results. Unlike the traditional fixed threshold warning system, it realizes the adjustment of the warning strategy according to the real-time data changes, which improves the reliability and response speed of the system. The warning unit compares the risk factor with the risk threshold, sets a multi-level warning level, distinguishes different degrees of risk, and avoids excessive warning or missed reports.

[0069] In some embodiments of the present application, when the acquisition unit determines the warning range according to the pipeline information and the obstacle information, it includes:

[0070] Pipeline information includes pipeline diameter, pipeline length, and pipeline quantity;

[0071] Obstacle information includes average plant density, average plant height, and the shortest distance between plants and pipelines;

[0072] ;

[0073] Where L represents the width of the warning range, Di represents the diameter of the i-th pipeline, Li represents the length of the i-th pipeline, N represents the number of pipelines, h represents the average height of the plants, and d represents the shortest distance between the plants and the pipelines. represents the average density of plants, k represents the distance attenuation coefficient, and the value range of k is [0.5, 0.9];

[0074] The warning range is formed by extending from the center of the pipeline to both sides of the pipeline, with the extension distance on each side being the width of the warning range.

[0075] It is understandable that the physical characteristics of the pipeline itself (such as diameter, length and number) are taken into account, and the obstacle information in the environment (such as the height, density and distance of the plants, etc.) is comprehensively considered, so that the determination of the warning range is more comprehensive and reflects the actual situation of the environment around the pipeline. The distance attenuation coefficient is adopted to automatically adjust the width of the warning range according to the relative distance between the plants and the pipeline in the actual environment. It has the ability of adaptive adjustment and dynamically adjusts the monitoring range according to the changes in the surrounding environment to ensure the comprehensiveness and timeliness of monitoring. It avoids the problem of unreasonable monitoring area setting in traditional monitoring methods and ensures the efficiency and accuracy of the monitoring system.

[0076] In some embodiments of the present application, when the judgment unit processes each sub-region data based on the Gaussian mixture model and judges whether the sub-region data changes, it includes:

[0077] The judgment unit collects historical image data, and selects image data having the same position as the sub-region data from the historical image data as model building data;

[0078] The judgment unit performs image segmentation on the model establishment data, removes the remaining images outside the pipeline, and initializes the mean, covariance matrix and mixing weight of the Gaussian mixture model with the pipeline image;

[0079] Based on the initialization parameter estimation model, the probability that each data point in the data belongs to each Gaussian distribution is established to obtain the posterior probability;

[0080] Update model parameters based on posterior probabilities;

[0081] The probability that each pixel of the sub-region data belongs to the current model is calculated, and when the likelihood of the pixel in the sub-region data is lower than the likelihood threshold, the judgment unit determines that the sub-region data has changed.

[0082] In some embodiments of the present application, calculating the probability that each pixel of the sub-region data belongs to the current model includes:

[0083] ;

[0084] Among them, P represents the likelihood of x data in the sub-region data, M represents the number of distributions in the Gaussian mixture model, represents the mixture weight of the j-th Gaussian distribution, Represents the probability density function of the j-th Gaussian distribution.

[0085] It can be understood that historical image data is first collected, which records the status of the pipeline and its surrounding environment. By screening the images in the historical images with the same data position as the current sub-area, data preparation samples are established for the subsequent Gaussian mixture model. Through image segmentation technology, the pipeline area in the image is separated from its surrounding environment. The parts outside the pipeline are eliminated and the focus is on the pipeline-related areas to reduce the computational burden and improve processing efficiency. At this time, the image data is mainly focused on the changes in the environment around the pipeline to ensure the accuracy of the model. The data in the pipeline area is used to initialize the parameters of the Gaussian mixture model:

[0086] Mean: The center of each Gaussian distribution, representing the expected value of the distribution.

[0087] Covariance Matrix: describes the dispersion and distribution of data and reflects the correlation between data.

[0088] Mixing Weights: Indicates the relative importance of each Gaussian distribution in the entire model, usually a training parameter of the model.

[0089] After initializing the parameters of the Gaussian mixture model, Bayesian inference is used to estimate the posterior probability of each data point (i.e., each pixel) belonging to each Gaussian distribution. The parameters of the Gaussian mixture model are updated through this probability, gradually improving the adaptability of the model. The calculation of the posterior probability is based on the known Gaussian distribution probability density function and the value of the current data point.

[0090] For each pixel, calculate the probability (likelihood) that it belongs to the current Gaussian mixture model. Determine how each pixel performs under the current model by comparing the degree of match between each pixel and different Gaussian distributions. If the likelihood of a pixel is lower than a set threshold (i.e., lower than the set likelihood threshold), the pixel is considered to have changed, indicating a change in the surrounding environment. When the likelihood of a pixel is detected to be lower than the set threshold, the environment of the pixel is considered to have changed. At this point, further analysis of the change is performed to determine whether it is a "biological change" or "abiotic change", such as changes in vegetation growth or artificial facilities.

[0091] It is understandable that the Gaussian mixture model can accurately capture the complex changes in the environment around the pipeline, especially in a changing environment. By continuously iteratively updating the model parameters, it can adapt to environmental changes and promptly detect potential risks. Based on image segmentation and Gaussian mixture model processing methods, it can efficiently identify the changed areas in a large amount of image data, especially distinguish between biological and non-biological changes, and provide accurate data support for the dynamic monitoring of the environment around the pipeline. By calculating the likelihood of each pixel and setting an appropriate threshold, the area where the change occurs can be identified. It avoids the problem of false alarms caused by excessive sensitivity and improves the reliability and stability of the system. By analyzing the image data within the warning range in real time and dynamically adjusting the monitoring strategy according to the current environmental changes, the system can not only cope with short-term changes in the environment around the pipeline, but also adapt to long-term environmental evolution.

[0092] In some embodiments of the present application, when the judgment unit identifies the change type and determines the risk factor according to all sub-regions where the change occurs, it includes:

[0093] The judgment unit performs image comparison on pixels that have changed in the sub-region data to determine the type of change, which includes biological change and non-biological change;

[0094] ;

[0095] Where A represents the area of ​​the changing region, Az represents the image area of ​​the sub-region, E represents the shortest distance between the changing region and the pipeline, E0 represents the safe distance of the pipeline, S represents the biological changing area, R0 represents the initial risk factor, and R0 is a positive integer.

[0096] In some embodiments of the present application, when the early warning unit compares the risk factor with the risk threshold and determines the early warning level, it includes: the early warning unit compares the risk factor with a first risk threshold and a second risk threshold that are preset respectively, and determines the early warning level according to the comparison results, and the first risk threshold is less than the second risk threshold;

[0097] Specifically, when the risk factor is less than or equal to the first risk threshold, the early warning unit determines the warning level as the first warning level; when the risk factor is greater than the first risk threshold and less than or equal to the second risk threshold, the early warning unit determines the warning level as the second warning level; when the risk factor is greater than the second risk threshold, the early warning unit determines the warning level as the third warning level; the first warning level is lower than the second warning level, and the second warning level is lower than the third warning level.

[0098] It is understandable that by calculating the area, distance and biological changes of the change area, different types of environmental changes can be identified, and potential risks can be judged based on the scale and nature of the changes. Dynamically adjust the risk assessment to avoid errors caused by static assessment. By setting multiple risk thresholds, it is ensured that different levels of warnings can be triggered at different risk levels, so that corresponding responses and decisions can be made according to different threat situations. It achieves flexible response to various changing environmental conditions, including biological and non-biological changes, and can better cope with the complexity of the external environment of the pipeline and improve pipeline safety.

[0099] In some embodiments of the present application, when the adjustment unit determines whether to adjust the risk factor according to the cumulative monitoring results, the adjustment unit includes: obtaining the area change curve of the change region according to the cumulative monitoring results, determining the maximum slope according to the area change curve, comparing the maximum slope with the slope threshold, and determining whether to adjust the risk factor according to the comparison result;

[0100] Specifically, when the maximum slope is greater than the slope threshold, the adjustment unit determines to adjust the risk factor; when the maximum slope is less than or equal to the slope threshold, the adjustment unit determines not to adjust the risk factor.

[0101] In some embodiments of the present application, when the adjustment unit determines to adjust the risk factor, it includes: the adjustment unit performs segment processing on the area change curve of the region;

[0102] Specifically, when there are peaks and valleys in the area change curve, the adjustment unit divides the regional area change curve into a first curve, a second curve and a third curve at the peak and valley positions; when there is only one of the peaks or valleys in the area change curve, the adjustment unit divides the regional area change curve into a first curve and a third curve at the peak or valley position; when the area change curve is an increasing function curve or a decreasing function curve, the adjustment unit divides the area change curve into a first curve and a third curve at the middle position of the area change curve; the adjustment unit adjusts the risk factor according to the first curve and the third curve to obtain the adjusted risk factor.

[0103] In some embodiments of the present application, the adjustment unit determines the adjustment factor according to the first curve and the third curve to adjust the risk factor, and obtains the adjusted risk factor, including:

[0104] The adjustment unit obtains the average change rate of the area of ​​the changing region according to the area change curve;

[0105] The adjustment factor is calculated as follows:

[0106]

[0107] in, represents the adjustment factor, T represents the number of area change rates of the changing regions in the first curve, Bt represents the area change rate of the t-th changing region in the first curve, B0 represents the average area change rate of the changing regions, Y represents the number of area change rates of the changing regions in the second curve, and By represents the area change rate of the y-th changing region in the second curve.

[0108] It is understandable that by analyzing the curve of the change in the area of ​​the variable area, the acceleration and deceleration stages of the change can be identified, and the risk factors can be adjusted based on these changes to improve the accuracy of risk prediction. The risk assessment can be adjusted in time according to the actual rate of environmental change to avoid static and lagging risk predictions, and better adapt to the complexity and dynamics of the environment around the pipeline. Through segmented processing and slope analysis, the response to changes in the environment around the pipeline can be refined, and different early warning responses can be made according to different stages of change. Not only can it cope with rapidly changing environments, but it can also adapt to gradually changing scenarios. By real-time monitoring and precise adjustment of risk factors, potential dangers in the environment around the pipeline can be identified in advance, reducing the risk of safety hazards such as damage, corrosion or leakage of the pipeline, thereby improving the safety and reliability of the pipeline.

[0109] In the above embodiments, by combining machine vision technology and Gaussian mixture model, automatic, real-time monitoring and risk assessment of the environment around the pipeline are realized. The acquisition unit locates the pipeline information and surrounding obstacles through edge recognition, realizes comprehensive perception of the monitoring range, and thus dynamically determines the warning range, effectively covering the environmental changes around the pipeline. The processing unit collects image data within the warning range in real time, and divides the data into sub-areas through image segmentation, which is convenient for more detailed analysis. The judgment unit uses the Gaussian mixture model to detect changes in each sub-area, automatically judges and identifies the type of change (such as biological change or non-biological change), and evaluates potential threats by calculating risk factors. The early warning unit judges the warning level based on the comparison between the risk factor and the preset threshold, and dynamically adjusts the monitoring strategy and warning level according to the actual situation, ensuring that it can respond quickly when changes occur and adjust the early warning measures in time. The accuracy and efficiency of monitoring the environment around the pipeline are improved, and potential safety hazards caused by human errors and environmental mutations are avoided.

[0110] In another preferred embodiment based on the above embodiment, refer to Figure 2 As shown, this embodiment provides a method for automatically monitoring the surrounding environment of a pipeline based on machine vision, which is applied to the above-mentioned automatic monitoring system for the surrounding environment of a pipeline based on machine vision, including:

[0111] S100: collecting image data within the range to be monitored, performing edge recognition on the image data, determining pipeline information and obstacle information, and determining a warning range based on the pipeline information and obstacle information;

[0112] S200: collecting image data of the warning range in real time as data to be analyzed, and dividing the data to be analyzed into a plurality of sub-area data;

[0113] S300: Processing each sub-region data based on the Gaussian mixture model and determining whether the sub-region data has changed; when it is determined that the sub-region data has changed, identifying the change type and determining the risk factor according to all the sub-regions that have changed;

[0114] S400: Compare the risk factor with the risk threshold to determine the warning level, which includes the first warning level, the second warning level and the third warning level.

[0115] S500: When the early warning unit determines that the early warning level is the second early warning level, the adjustment unit performs cumulative monitoring on the sub-area where the change occurs, determines whether to adjust the risk factor according to the cumulative monitoring result, and updates the early warning level according to the adjusted risk factor.

[0116] It is understandable that by combining machine vision technology and Gaussian mixture model, the automation, real-time monitoring and risk assessment of the pipeline surrounding environment are realized. The acquisition unit locates the pipeline information and surrounding obstacles through edge recognition, realizes comprehensive perception of the monitoring range, and dynamically determines the warning range to effectively cover the environmental changes around the pipeline. The processing unit collects image data within the warning range in real time, and divides the data into sub-areas through image segmentation for more detailed analysis. The judgment unit uses the Gaussian mixture model to detect changes in each sub-area, automatically judges and identifies the type of change (such as biological change or non-biological change), and evaluates potential threats by calculating risk factors. The early warning unit judges the warning level based on the comparison between the risk factor and the preset threshold, and dynamically adjusts the monitoring strategy and warning level according to the actual situation, ensuring that it can respond quickly when changes occur and adjust the early warning measures in time. The accuracy and efficiency of monitoring the environment around the pipeline are improved, and potential safety hazards caused by human errors and environmental mutations are avoided.

[0117] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0118] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems) and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0119] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0120] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0121] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A pipeline surrounding environment automatic monitoring system based on machine vision, characterized in that: include: A collection unit is configured to collect image data within a range to be monitored, perform edge recognition on the image data, determine pipeline information and obstacle information, and determine a warning range according to the pipeline information and obstacle information; A processing unit is configured to collect image data of the warning range in real time as data to be analyzed, and divide the data to be analyzed into a plurality of sub-area data; a judgment unit configured to process each of the sub-region data based on a Gaussian mixture model and judge whether the sub-region data has changed; when the judgment unit determines that the sub-region data has changed, identify the change type and determine the risk factor according to all the sub-regions that have changed; an early warning unit, configured to compare the risk factor with a risk threshold to determine an early warning level, wherein the early warning level includes a first early warning level, a second early warning level, and a third early warning level; The adjustment unit is configured to, when the early warning unit determines that the early warning level is the second early warning level, perform cumulative monitoring on the sub-area where the change occurs, and determine whether to adjust the risk factor based on the cumulative monitoring results, and update the early warning level based on the adjusted risk factor.

2. The automatic monitoring system for pipeline surrounding environment based on machine vision according to claim 1 is characterized in that: When the acquisition unit determines the warning range according to the pipeline information and the obstacle information, it includes: The pipeline information includes pipeline diameter, pipeline length and pipeline quantity; The obstacle information includes the average density of plants, the average height of plants, and the shortest distance between plants and pipelines; ； Where L represents the width of the warning range, Di represents the diameter of the i-th pipeline, Li represents the length of the i-th pipeline, N represents the number of pipelines, h represents the average height of the plants, and d represents the shortest distance between the plants and the pipelines. represents the average density of plants, k represents the distance attenuation coefficient, and the value range of k is [0.5, 0.9]; The warning range is formed by extending from the center of the pipeline to both sides of the pipeline, with the extension distance on each side being the width of the warning range.

3. The automatic monitoring system for pipeline surrounding environment based on machine vision according to claim 1 is characterized in that: The determining unit processes each of the sub-region data based on the Gaussian mixture model and determines whether the sub-region data changes, including: The judgment unit collects historical image data, and selects image data having the same position as the sub-region data from the historical image data as model building data; The judgment unit performs image segmentation on the model establishment data, removes the remaining images outside the pipeline, and initializes the mean, covariance matrix and mixing weight of the Gaussian mixture model with the pipeline image; Estimate the probability that each data point in the model establishment data belongs to each Gaussian distribution based on the initialization parameters to obtain the posterior probability; updating model parameters according to the posterior probability; The probability that each pixel of the sub-region data belongs to the current model is calculated, and when the likelihood of the pixel in the sub-region data is lower than a likelihood threshold, the judgment unit determines that the sub-region data has changed.

4. The automatic monitoring system for pipeline surrounding environment based on machine vision according to claim 3 is characterized in that: Calculating the probability that each pixel of the sub-region data belongs to the current model includes: ； Among them, P represents the likelihood of x data in the sub-region data, M represents the number of distributions in the Gaussian mixture model, represents the mixture weight of the j-th Gaussian distribution, Represents the probability density function of the j-th Gaussian distribution.

5. The automatic monitoring system for pipeline surrounding environment based on machine vision according to claim 4 is characterized in that: When the judgment unit identifies the change type and determines the risk factor according to all the sub-areas where the change occurs, it includes: The judgment unit performs image comparison on pixels that have changed in the sub-region data to determine the change type, where the change type includes biological change and non-biological change; ； Where A represents the area of ​​the changing region, Az represents the image area of ​​the sub-region, E represents the shortest distance between the changing region and the pipeline, E0 represents the safe distance of the pipeline, S represents the biological changing area, R0 represents the initial risk factor, and R0 is a positive integer.

6. The automatic monitoring system for pipeline surrounding environment based on machine vision according to claim 1 is characterized in that: When the early warning unit compares the risk factor with the risk threshold and determines the early warning level, it includes: The early warning unit compares the risk factor with a first risk threshold and a second risk threshold set in advance, respectively, and determines the early warning level according to the comparison result, wherein the first risk threshold is less than the second risk threshold; When the risk factor is less than or equal to a first risk threshold, the warning unit determines the warning level as the first warning level; when the risk factor is greater than the first risk threshold and less than or equal to a second risk threshold, the warning unit determines the warning level as the second warning level; when the risk factor is greater than the second risk threshold, the warning unit determines the warning level as the third warning level; the first warning level is lower than the second warning level, and the second warning level is lower than the third warning level.

7. The automatic monitoring system for pipeline surrounding environment based on machine vision according to claim 6 is characterized in that: When the adjustment unit determines whether to adjust the risk factor according to the accumulated monitoring results, it includes: The adjustment unit obtains an area change curve of the change region according to the accumulated monitoring results, determines a maximum slope according to the area change curve, compares the maximum slope with a slope threshold, and determines whether to adjust the risk factor according to the comparison result; When the maximum slope is greater than the slope threshold, the adjustment unit determines to adjust the risk factor; When the maximum slope is less than or equal to a slope threshold, the adjustment unit determines not to adjust the risk factor.

8. The automatic monitoring system for pipeline surrounding environment based on machine vision according to claim 7 is characterized in that: When the adjustment unit determines to adjust the risk factor, it includes: The adjustment unit performs segment processing on the region area change curve; When the area change curve has a peak and a valley point, the adjustment unit divides the area change curve into a first curve, a second curve and a third curve at the peak and the valley point positions; When there is only one of a peak point or a valley point in the area change curve, the adjustment unit divides the area change curve into a first curve and a third curve at the peak point or the valley point; When the area change curve is an increasing function curve or a decreasing function curve, the adjustment unit divides the area change curve into a first curve and a third curve at a middle position of the area change curve; The adjustment unit adjusts the risk factor according to the first curve and the third curve to obtain an adjusted risk factor.

9. The automatic monitoring system for pipeline surrounding environment based on machine vision according to claim 8 is characterized in that: The adjusting unit determines the adjustment factor according to the first curve and the third curve to adjust the risk factor, and obtains the adjusted risk factor, including: The adjusting unit obtains an average rate of change of the area of ​​the changing region according to the area change curve; The adjustment factor is calculated by the following formula: ； in, represents the adjustment factor, T represents the number of area change rates of the changing regions in the first curve, Bt represents the area change rate of the t-th changing region in the first curve, B0 represents the average area change rate of the changing regions, Y represents the number of area change rates of the changing regions in the second curve, and By represents the area change rate of the y-th changing region in the second curve.

10. A method for automatically monitoring the surrounding environment of a pipeline based on machine vision, applied to the automatic monitoring system for the surrounding environment of a pipeline based on machine vision as claimed in any one of claims 1 to 9, characterized in that: include: Collect image data within the range to be monitored, perform edge recognition on the image data, determine pipeline information and obstacle information, and determine the warning range based on the pipeline information and obstacle information; Collecting image data of the warning range in real time as data to be analyzed, and dividing the data to be analyzed into a plurality of sub-area data; Processing each of the sub-region data based on a Gaussian mixture model and determining whether the sub-region data has changed; when it is determined that the sub-region data has changed, identifying the type of change and determining a risk factor based on all the sub-regions that have changed; Comparing the risk factor with the risk threshold to determine a warning level, wherein the warning level includes a first warning level, a second warning level, and a third warning level; When the warning unit determines that the warning level is the second warning level, the adjustment unit performs cumulative monitoring on the sub-area where the change occurs, determines whether to adjust the risk factor according to the cumulative monitoring result, and updates the warning level according to the adjusted risk factor.

Images