Three-dimensional space sensing tower electronic fence device and control method
Through infrared imaging and video analysis technology combined with three-dimensional modeling, three-dimensional spatial perception of the area around the power tower is achieved, solving the problem that traditional two-dimensional image processing is difficult to monitor the safety of the power tower, and improving the safety monitoring accuracy and efficiency of the power tower. Provides guarantee for the safe operation of the power grid.
Patent Information
- Application Number
- CN202411124985.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-16
- Publication Date
- 2024-11-22
AI Technical Summary
Traditional electronic fence systems are based on two-dimensional image processing, making it difficult to accurately perceive the three-dimensional spatial information around the power tower, resulting in limited monitoring effects.
Using infrared imaging technology, video analysis technology and three-dimensional modeling technology, the infrared image and video data around the power tower are monitored, video features and perception data are extracted, target tracking and three-dimensional spatial perception model are updated, and the area around the power tower is realized. three-dimensional spatial perception and monitoring.
It improves the safety monitoring accuracy and efficiency of the power tower, provides strong guarantees for the safe operation of the power grid, can accurately detect the size, shape and motion trajectory of the intruded target, and enhances the real-time monitoring and response capabilities for illegal intrusions.
Smart Images

Figure CN119007357A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of electric power grid, and in particular to a three-dimensional space sensing pole tower electronic fence device and a control method. Background Art
[0002] With the rapid development of smart grids, power towers are important infrastructure for power grids, and their safety and stability are crucial to the normal operation of the power system. However, power towers are usually distributed in the wild or remote areas, and are vulnerable to security threats such as illegal intrusion, theft, and destruction. Traditional security monitoring methods, such as manual inspections and fixed camera monitoring, have problems such as large monitoring blind spots, slow response speeds, and high false alarm rates, which are difficult to meet the security monitoring needs of modern power grids.
[0003] In order to improve the safety monitoring capabilities of power towers, electronic fence technology has been widely used in the safety protection of power towers in recent years. Electronic fences set up virtual or physical fences around power towers, combined with sensors, cameras and other equipment, to achieve real-time monitoring and alarm of intrusion behaviors in the fenced area. However, traditional electronic fence systems are mostly based on two-dimensional image processing, and it is difficult to accurately perceive the three-dimensional spatial information of the intrusion target, such as the target's height, depth, movement trajectory, etc., resulting in limited monitoring effects. Summary of the invention
[0004] The present invention provides a three-dimensional space perception pole tower electronic fence device and control method, which can realize three-dimensional space perception and monitoring of the area around the power pole tower through infrared imaging technology, video analysis technology and three-dimensional modeling technology, thereby improving the safety monitoring accuracy and efficiency of the power pole tower and providing a strong guarantee for the safe operation of the power grid.
[0005] In a first aspect, the present invention provides a control method for a three-dimensional space perception pole tower electronic fence device, the method comprising: monitoring infrared images of an electronic fence area around a power pole tower; analyzing the infrared images to determine whether there is an illegal intrusion in the area around the power pole tower; if there is an illegal intrusion, extracting video data of a target area where the illegal intrusion occurs, and analyzing the video data frame by frame to determine video features of the target area, the video features including video features of each frame of the video data, differential features between two consecutive frames of images, and cumulative differential features; based on the video features of the target area, tracking the target to determine perception data of the target area, the perception data including geometric data and trajectory data of the intruding target; based on the perception data of the target area, updating a three-dimensional space perception model of the power pole tower, and visually displaying the three-dimensional space perception model.
[0006] In a possible implementation, an infrared image is analyzed to determine whether there is any illegal intrusion in the area around the power pole tower, including: for any area in the area around the power pole tower, grayscale calculation is performed based on a standard image of the area to determine the grayscale threshold of each pixel in the area; the standard image is an image of the area when there is no intrusion; the infrared image is converted into a grayscale image, and the grayscale value of each pixel in the grayscale image is calculated; based on the grayscale value of each pixel in the grayscale image and the grayscale threshold of each pixel, threshold judgment is performed to determine abnormal pixels in the area; based on the abnormal pixels in the area, the image block where the intrusion target is located is determined; based on the infrared image and the image block where the intrusion target is located, edge detection is performed to determine the geometric data and position information of the intrusion target; the geometric data includes the size and shape of the intrusion target, and the position information is used to characterize the relative position relationship between the center of the intrusion target and the electronic fence; based on the geometric data and position information of the intrusion target, it is determined whether there is illegal intrusion in the area.
[0007] In a possible implementation, video data is analyzed frame by frame to determine video features of a target area, including: performing frame processing on video data to obtain multiple frames of video images; performing feature extraction on each frame of video images to determine global features and local features of each frame of video images; calculating pixel differences between two consecutive frames of video images in the multiple frames of video images to generate multiple frames of differential images; performing feature extraction on the multiple frames of differential images to determine differential features, which include global differential features and local differential features; accumulating the multiple frames of differential images to obtain a cumulative differential image; performing feature extraction on the cumulative differential images to determine cumulative differential features of the cumulative differential images, which include global cumulative features and local cumulative features; and generating video features of the target area based on the global features and local features, differential features, and cumulative differential features of each frame of video images.
[0008] In a possible implementation, feature extraction is performed on each frame of video image to determine the global features and local features of each frame of video image, including: for any frame of video image, the global features of the frame of video image are calculated, and the global features include the histogram and color moment of the video image; the histogram is used to characterize the probability of occurrence of each grayscale level on the grayscale image of the video image; the color moment is used to characterize the distribution of each color on the video image; traverse each pixel point of the frame of video image, take each pixel point as the center, take the first threshold number of pixels as the radius, and determine the circumference corresponding to each pixel point; compare the brightness of each pixel point on the circumference corresponding to each pixel point with the brightness of the central pixel point to determine the comparison result; if the comparison result meets the set condition, the central pixel point is determined as a candidate point, and the preset condition is that the continuous pixel point on the circumference The difference between the brightness of the second threshold number of pixel points and the brightness of the candidate point is greater than the set brightness difference; for multiple adjacent candidate points, non-maximum suppression is adopted to delete the candidate points in the multiple adjacent candidate points that are non-maximum, and retain the candidate points in the maximum as key points to form a key point group; based on the position of each key point in the key point group in the frame video image, the grayscale centroid of the key point group is calculated; based on the grayscale centroid of the key point group and the position of each key point, the main direction of each key point is determined; along the main direction of each key point, the brightness of each key point in the main direction is compared to determine the flag bit of each key point; the flag bit is used to characterize the brightness change of the key point along the main direction; based on the position, brightness, grayscale centroid and flag bit of each key point on the frame video image, the local features of the frame video image are determined.
[0009] In a possible implementation, target tracking is performed based on video features of the target area to determine perception data of the target area, including: inputting the video features of the target area into a preset target tracking model to obtain feature data of the intrusion target, the feature data of the intrusion target including geometric data of the intrusion target and position information of the intrusion target on each frame of the video image; determining the actual position of the intrusion target in the target area based on the position information of the intrusion target on each frame of the video image; determining the trajectory data of the intrusion target based on the actual position of the intrusion target in the target area and the time corresponding to each frame of the image.
[0010] In a possible implementation, target tracking is performed based on the video features of the target area, and before the perception data of the target area is determined, the method also includes: obtaining video images of various intruders in motion, as well as geometric data and position information of various intruders; performing frame-by-frame analysis based on the video images of various intruders in motion to generate video features of various intruders; generating training samples using the video features of various intruders as input and the geometric data and position information of various intruders as output; and training a neural network based on the training samples to obtain a target tracking model.
[0011] In one possible implementation, a neural network is trained based on training samples to obtain a target tracking model, including: training a neural network based on the training samples to obtain an initial model; taking the initial model as a teacher model, and constructing a student model with the same model parameters as the teacher model; testing the teacher model based on a test set in the training samples to obtain a first test accuracy of the teacher model; determining a test accuracy threshold based on the first test accuracy; performing knowledge distillation on the student model to obtain a distilled student model; testing the distilled student model based on the test set in the training samples to obtain a second test accuracy; if the second test accuracy is greater than or equal to the test accuracy threshold, repeatedly performing knowledge distillation on the student model until the second test accuracy is less than the test accuracy threshold; if the second test accuracy is less than the test accuracy threshold, using the student model before the knowledge distillation of the current student model as the target tracking model.
[0012] In one possible implementation, the three-dimensional spatial perception model of the power pole tower is updated based on the perception data of the target area, including: performing three-dimensional modeling based on the geometric data of the invading target on each frame of the video image to determine the three-dimensional model of the invading target; updating the three-dimensional spatial perception model of the power pole tower based on the three-dimensional model of the invading target and the trajectory data of the invading target to obtain the three-dimensional spatial perception model at each moment.
[0013] In a possible implementation, after performing target tracking based on the video features of the target area and determining the perception data of the target area, the method further includes: judging an intruder based on the perception data and determining the intruder information, wherein the intruder information includes the type and quantity; querying a database based on the intruder information and generating an expulsion control instruction; the expulsion control instruction is used to expel the intruder from the target area.
[0014] In a possible implementation, the method further includes: acquiring infrared images and video data of power towers, power transmission equipment on power towers, and power transmission lines on power towers; determining temperature distribution of power towers, power transmission equipment on power towers, and power transmission lines on power towers based on the infrared images; determining appearance information of power towers, power transmission equipment on power towers, and power transmission lines on power towers based on the video data; the appearance information includes whether they are damaged, deformed, and whether there are foreign objects attached; and performing fault detection based on the temperature distribution and appearance information to determine the working status of power towers, power transmission equipment, and transmission lines.
[0015] In a second aspect, an embodiment of the present invention provides a control device for a three-dimensional space perception pole tower electronic fence device, the control device comprising: a communication module for monitoring infrared images of the electronic fence area around the power pole tower; a processing module for analyzing the infrared image to determine whether there is illegal intrusion in the area around the power pole tower; if there is illegal intrusion, extracting video data of the target area where the illegal intrusion occurs, and analyzing the video data frame by frame to determine the video features of the target area, the video features including the video features of each frame of the video data, the differential features between two consecutive frames of images, and the cumulative differential features; based on the video features of the target area, tracking the target, and determining the perception data of the target area, the perception data including the geometric data and trajectory data of the intruding target; based on the perception data of the target area, updating the three-dimensional space perception model of the power pole tower, and visually displaying the three-dimensional space perception model.
[0016] In the third aspect, an embodiment of the present invention provides a three-dimensional space perception pole tower electronic fence device, which includes an infrared detection device, a camera device and a control device, and the control device is used to execute the steps of the method described in the above-mentioned first aspect and any possible implementation method of the first aspect.
[0017] The present invention provides a three-dimensional space perception pole tower electronic fence device and control method. The present invention determines whether there is illegal intrusion through infrared detection, and performs video analysis on the intrusion area to obtain the video features, differential features and cumulative differential features of the intruder, comprehensively analyze the motion characteristics of the intrusion target, realize the tracking of the intrusion target, and update the three-dimensional space perception model for visual display. The present invention realizes three-dimensional space perception and monitoring of the area around the power pole tower through infrared imaging technology, video analysis technology and three-dimensional modeling technology, improves the safety monitoring accuracy and efficiency of the power pole tower, and provides a strong guarantee for the safe operation of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS
[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.
[0019] Figure 1 It is a flow chart of a control method of a three-dimensional space sensing pole tower electronic fence device provided by an embodiment of the present invention;
[0020] Figure 2 It is a structural schematic diagram of a control device of a three-dimensional space perception pole tower electronic fence device provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0021] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present invention. However, it should be clear to those skilled in the art that the present invention may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present invention.
[0022] In the description of the present invention, unless otherwise specified, " / " means "or", for example, A / B can mean A or B. "And / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. In addition, "at least one" and "plurality" refer to two or more. The words "first", "second", etc. do not limit the quantity and execution order, and the words "first", "second", etc. do not limit them to be different.
[0023] In the embodiments of the present application, words such as "exemplary" or "for example" are used to indicate examples, illustrations or descriptions. Any embodiment or design described as "exemplary" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary" or "for example" is intended to present related concepts in a concrete way for easy understanding.
[0024] In addition, the terms "including" and "having" and any variations thereof mentioned in the description of this application are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or modules is not limited to the listed steps or modules, but may optionally include other steps or modules that are not listed, or may optionally include other steps or modules that are inherent to these processes, methods, products or devices.
[0025] In order to make the purpose, technical solutions and advantages of the present invention more clear, the following will be described through specific embodiments in conjunction with the accompanying drawings of the present invention.
[0026] As described in the background art, traditional electronic fence systems are mostly based on two-dimensional image processing, which makes it difficult to accurately perceive the three-dimensional spatial information of the intrusion target, and there is a technical problem of limited monitoring effect.
[0027] To solve this technical problem, Figure 1 As shown, an embodiment of the present invention provides a control method for a three-dimensional space sensing pole tower electronic fence device. The method includes steps S101-S105.
[0028] S101, monitoring infrared images of the electronic fence area around the power tower.
[0029] In some embodiments, infrared detection equipment and camera equipment are installed around the power pole tower.
[0030] Exemplarily, the infrared detection equipment and the camera equipment may be installed on the electronic fence around the power pole tower to capture infrared images and video data around the electronic fence.
[0031] As another example, the infrared detection equipment and the camera equipment may be installed on the power pole tower to capture infrared images and video data around the power pole tower.
[0032] It should be noted that the intrusion targets around power poles can be people, animals, robots, special equipment, etc. When people and animals enter the area around power poles, infrared images are easier to detect because people and animals are more sensitive to infrared rays. In addition, for robots and special equipment, since they are equipped with energy devices and generate heat, infrared images for detecting intrusions are easier to detect.
[0033] When an intrusion is detected, the embodiment of the present invention can combine infrared detection technology and video analysis technology to determine the size, shape, trajectory and other information of the intruder, and then combine it with a three-dimensional space perception model for visualization, so as to facilitate users to observe the intruder.
[0034] S102: Analyze the infrared image to determine whether there is any illegal intrusion in the area around the power tower.
[0035] As a possible implementation manner, step S102 may be specifically implemented as steps S1021 - S1026 .
[0036] S1021. For any area in the area surrounding the power pole tower, grayscale calculation is performed based on a standard image of the area to determine a grayscale threshold of each pixel in the area.
[0037] In some embodiments, the standard image is an image when the area is in a non-invasion state.
[0038] Exemplarily, the embodiment of the present invention can directly determine the gray value of each pixel in the standard image as the gray threshold. Alternatively, the embodiment of the present invention can also determine the product of the gray value of each pixel in the standard image and the correlation coefficient as the gray threshold.
[0039] S1022, converting the infrared image into a grayscale image, and calculating the grayscale value of each pixel in the grayscale image.
[0040] S1023: Based on the grayscale value of each pixel in the grayscale image and the grayscale threshold of each pixel, threshold judgment is performed to determine abnormal pixels in the area.
[0041] Exemplarily, the embodiment of the present invention can calculate the difference between the gray value of each pixel in the gray image and the gray threshold of the corresponding pixel, and if the difference is greater than the set threshold, it is determined to be an abnormal pixel. If the difference is less than the set threshold, it is determined to be a normal pixel.
[0042] S1024: Determine the image block where the intrusion target is located based on the abnormal pixels in the area.
[0043] Exemplarily, the embodiment of the present invention may divide a video image into a plurality of image blocks, and determine the image block with abnormal pixels as the image block where the intrusion target is located.
[0044] S1025: Perform edge detection based on the infrared image and the image block where the intrusion target is located to determine the geometric data and position information of the intrusion target.
[0045] In some embodiments, the geometric data includes the size and shape of the intrusion target. The position information is used to characterize the relative position relationship between the center of the intrusion target and the electronic fence.
[0046] Exemplarily, an embodiment of the present invention can perform edge detection based on an infrared image and an image block where the intrusion target is located to determine the boundary of the intrusion target, determine the geometric data of the intrusion target based on the boundary of the intrusion target, determine the center of the intrusion target and the distance between the center and an electronic fence based on the geometric data of the intrusion target; determine the location information of the intrusion target based on the geometric data of the intrusion target, the center of the intrusion target and the distance between the center and the electronic fence.
[0047] S1026. Determine whether there is any illegal intrusion in the area based on the geometric data and location information of the intrusion target.
[0048] Exemplarily, the embodiment of the present invention can determine whether the intrusion target is in the alarm area based on the geometric data and location information of the intrusion target. If the intrusion target is in the alarm area, it is determined that there is illegal intrusion in the area. If the intrusion target is not in the alarm area, it is determined that there is no illegal intrusion in the area. The alarm area is an area in the surrounding area of the power pole tower where intrusion is prohibited.
[0049] S103: If there is illegal intrusion, extract the video data of the target area where the illegal intrusion occurs, and analyze the video data frame by frame to determine the video features of the target area.
[0050] In the embodiment of the present application, the video features include the video features of each frame image in the video data, the differential features between two consecutive frame images, and the cumulative differential features.
[0051] As a possible implementation manner, step S103 may be specifically implemented as steps S1031 - S1037 .
[0052] S1031 , performing frame processing on the video data to obtain multiple frames of video images.
[0053] S1032: Extract features from each frame of video image to determine global features and local features of each frame of video image.
[0054] Exemplarily, step S1032 may be specifically implemented as steps A1-A9.
[0055] A1. For any frame of video image, calculate the global features of the frame of video image, where the global features include the histogram and color moment of the video image.
[0056] In some embodiments, the histogram is used to characterize the probability of occurrence of each grayscale level on the grayscale image of the video image. The color moment is used to characterize the distribution of each color on the video image.
[0057] Exemplarily, the embodiment of the present invention can count the number of pixels corresponding to each gray level, determine the probability of each gray level appearing based on the number of pixels corresponding to each gray level and the total number of pixels, and draw a histogram based on the probability of each gray level appearing. In addition, the embodiment of the present invention can also construct a histogram vector based on the probability of each gray level appearing, so as to facilitate the combination to obtain video features.
[0058] Exemplarily, the embodiment of the present invention can calculate the first-order moment, second-order moment and third-order moment in the color moment, realize the calculation of the color moment, and construct the color moment vector. Among them, the first-order moment is the color average value. The embodiment of the present invention can calculate the average value of all pixel values of each color channel in multiple color channels, and determine the average value as the color average value, that is, the first-order moment. The first-order moment is used to reflect the overall trend of the image color.
[0059] The embodiment of the present invention can also calculate the difference between the pixel value in each color channel and the color average value, and calculate the average value of the square of the difference to determine the color variance, that is, the second-order moment. The second-order moment is used to reflect the discrete degree of image color.
[0060] The embodiment of the present invention can also calculate the average value of the cube of the difference between the pixel value of each color channel and the color average value, and then divide it by the cube of the color variance to determine the color skewness, that is, the third-order moment. The third-order moment is used to reflect the degree of skewness of the image color distribution.
[0061] A2. Traverse each pixel point of the frame of video image, take each pixel point as the center, take the first threshold number of pixels points as the radius, and determine the circumference corresponding to each pixel point.
[0062] A3. Compare the brightness of each pixel point on the circumference corresponding to each pixel point with the brightness of the central pixel point to determine the comparison result.
[0063] A4. If the comparison result meets the set conditions, the central pixel point is determined as the candidate point.
[0064] In some embodiments, the preset condition is that the difference between the brightness of a second threshold number of consecutive pixel points on the circumference and the brightness of the candidate point is greater than a set brightness difference.
[0065] A5. For multiple adjacent candidate points, non-maximum suppression is used to delete the candidate points with non-maximum values among the multiple adjacent candidate points, and the candidate points with maximum values are retained as key points to form a key point group.
[0066] A6. Calculate the grayscale centroid of the key point group based on the position of each key point in the key point group in the frame of video image.
[0067] A7. Based on the grayscale centroid of the key point group and the position of each key point, determine the main direction of each key point.
[0068] A8. Along the main direction of each key point, compare the brightness of each key point in the main direction to determine the marker position of each key point.
[0069] In some embodiments, the flag bit is used to characterize the brightness change of the key point along the main direction.
[0070] A9. Determine the local features of the frame of video image based on the position, brightness, grayscale centroid and flag of each key point on the frame of video image.
[0071] S1033: Calculate the pixel difference between two consecutive video frames in the multi-frame video image to generate a multi-frame differential image.
[0072] S1034: extract features from multiple frames of differential images to determine differential features.
[0073] In some embodiments, the differential features include global differential features and local differential features.
[0074] S1035 , accumulating multiple frames of differential images to obtain an accumulated differential image.
[0075] S1036: Extract features from the cumulative difference image to determine the cumulative difference features of the cumulative difference image.
[0076] In some embodiments, the accumulated differential features include global accumulated features and local accumulated features.
[0077] S1037. Generate video features of the target area based on the global features and local features, differential features, and accumulated differential features of each frame of video image.
[0078] In this way, the embodiment of the present invention can comprehensively consider the video features, differential features and cumulative differential features of each frame of video image, that is, consider the impact of the speed of movement of the intrusion target, so that the detection of the intrusion target is more precise and accurate, and the detection efficiency of the intrusion target is improved.
[0079] S104: Track the target based on the video features of the target area to determine the perception data of the target area.
[0080] In an embodiment of the present application, the perception data includes geometric data and trajectory data of the intrusion target.
[0081] As a possible implementation manner, step S104 may be specifically implemented as steps S1041 - S1045 .
[0082] S1041. Input the video features of the target area into a preset target tracking model to obtain feature data of the intrusion target.
[0083] In some embodiments, the feature data of the intrusion target includes geometric data of the intrusion target and position information of the intrusion target on each frame of the video image.
[0084] S1042: Determine the actual position of the intrusion target in the target area based on the position information of the intrusion target in each frame of the video image.
[0085] S1043: Determine the trajectory data of the intrusion target based on the actual position of the intrusion target in the target area and the time corresponding to each frame of the image.
[0086] S105. Based on the perception data of the target area, update the three-dimensional space perception model of the power tower, and visualize the three-dimensional space perception model.
[0087] As a possible implementation manner, step S105 can be specifically implemented as steps S1051 - S1052 .
[0088] S1051. Perform three-dimensional modeling based on the geometric data of the intrusion target in each frame of video image to determine the three-dimensional model of the intrusion target.
[0089] S1052: Based on the three-dimensional model of the intrusion target and the trajectory data of the intrusion target, the three-dimensional space perception model of the power tower is updated to obtain the three-dimensional space perception model at each moment.
[0090] The present invention provides a three-dimensional space perception pole tower electronic fence device and control method, which determines whether there is illegal intrusion through infrared detection, and performs video analysis on the intrusion area to obtain the video features, differential features and cumulative differential features of the intruder, comprehensively analyze the motion characteristics of the intrusion target, realize the tracking of the intrusion target, and update the three-dimensional space perception model for visual display. The present invention realizes three-dimensional space perception and monitoring of the area around the power pole tower through infrared imaging technology, video analysis technology and three-dimensional modeling technology, improves the safety monitoring accuracy and efficiency of the power pole tower, and provides a strong guarantee for the safe operation of the power grid.
[0091] The present invention realizes three-dimensional spatial perception and monitoring of the area around the power tower by integrating advanced technical means such as infrared imaging technology, video analysis technology, and three-dimensional modeling technology. Specifically, the method monitors the infrared images around the power tower and analyzes the infrared images using image processing algorithms to determine whether there is illegal intrusion. Once an illegal intrusion is detected, the video data of the target area is immediately extracted, and frame-by-frame analysis is performed to extract video features. Based on these video features, the target tracking algorithm is used to determine the geometric data and trajectory data of the intrusion target, and then the three-dimensional spatial perception model of the power tower is updated and visualized. This method not only improves the accuracy and efficiency of safety monitoring of power towers, but also realizes three-dimensional spatial tracking and visual display of intrusion behaviors, providing a strong guarantee for the safe operation of the power grid.
[0092] Optionally, the control method of the three-dimensional space perception pole tower electronic fence device provided by the embodiment of the present invention further includes steps S201-S204 before step S104.
[0093] S201, obtaining video images of various intrusion objects during their movement, as well as geometric data and position information of various intrusion objects.
[0094] S202: Based on the video images of various intruders in motion, frame-by-frame analysis is performed to generate video features of various intruders.
[0095] S203, using the video features of various intrusion objects as input and the geometric data and position information of various intrusion objects as output to generate training samples.
[0096] S204: Perform neural network training based on the training samples to obtain a target tracking model.
[0097] As a possible implementation manner, step S204 can be specifically implemented as steps S2041-S2047.
[0098] S2041. Perform neural network training based on the training samples to obtain an initial model.
[0099] S2042. Take the initial model as the teacher model and construct a student model with the same model parameters as the teacher model.
[0100] S2043. Test the teacher model based on the test set in the training sample to obtain a first test accuracy of the teacher model.
[0101] S2044. Determine a test accuracy threshold based on the first test accuracy.
[0102] S2045. Perform knowledge distillation on the student model to obtain a distilled student model.
[0103] S2046. Test the distilled student model based on the test set in the training sample to obtain a second test accuracy.
[0104] S2047. If the second test accuracy is greater than or equal to the test accuracy threshold, repeatedly perform knowledge distillation on the student model until the second test accuracy is less than the test accuracy threshold.
[0105] S2048. If the second test accuracy is less than the test accuracy threshold, the student model before the knowledge distillation of the current student model is used as the target tracking model.
[0106] In this way, the embodiment of the present invention can obtain a target tracking model through neural network training before performing video analysis and target tracking on the intrusion target, which is convenient for detecting and judging the intrusion target. In addition, during the training process, the model is subjected to knowledge distillation, which simplifies the model complexity and improves the calculation speed of the model on the basis of ensuring the test accuracy, thereby ensuring the real-time detection of intrusion targets on power towers.
[0107] Optionally, the control method of the three-dimensional space perception pole tower electronic fence device provided by the embodiment of the present invention further includes steps S301-S302 after step S104.
[0108] S301. Based on the perception data, make an intrusion judgment and determine the intrusion information.
[0109] In some embodiments, the intruder information includes type and quantity.
[0110] S302: Based on the intruder information, query the database and generate an expulsion control instruction.
[0111] In some embodiments, the expulsion control instructions are used to expel intruders from a target area.
[0112] In this way, the embodiment of the present invention can judge the intruder by the size, shape and movement trajectory of the intruder in the sensing data, determine the number and types of the intruder, formulate corresponding expulsion instructions, expel the intruder, and ensure the safe operation of the power pole tower related equipment.
[0113] Optionally, the control method of the three-dimensional space perception pole tower electronic fence device provided by the embodiment of the present invention also includes steps S401-S402.
[0114] S401, acquiring infrared images and video data of power towers, power transmission equipment on the power towers, and power transmission lines on the power towers.
[0115] S402: Based on the infrared image, determine the temperature distribution of the power tower, the power transmission equipment on the power tower, and the power transmission line on the power tower.
[0116] S403: Determine appearance information of the power tower, power transmission equipment on the power tower, and power transmission lines on the power tower based on the video data.
[0117] In some embodiments, the appearance information includes whether it is damaged, deformed, and whether there is foreign matter attached.
[0118] S404: Based on the temperature distribution and appearance information, perform fault detection to determine the working status of the power tower, transmission equipment and transmission line.
[0119] In this way, the embodiment of the present invention can perform infrared scanning and video scanning on power poles and related equipment and lines, detect possible faults in power poles and towers from the two aspects of heat generation and appearance detection, determine the working status of power poles, transmission equipment and transmission lines, which is beneficial to the inspection of power poles and towers and improves the inspection efficiency.
[0120] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.
[0121] The following is an embodiment of the device of the present invention. For details not described in detail therein, reference may be made to the corresponding method embodiment described above.
[0122] Figure 2 The schematic diagram of the structure of a control device of a three-dimensional space sensing pole tower electronic fence device provided by an embodiment of the present invention is shown. The control device 500 includes a communication module 501 and a processing module 502 .
[0123] The communication module 501 is used to monitor the infrared image of the electronic fence area around the power tower.
[0124] The processing module 502 is used to analyze the infrared image to determine whether there is any illegal intrusion in the area around the power tower; if there is any illegal intrusion, the video data of the target area where the illegal intrusion occurs is extracted, and the video data is analyzed frame by frame to determine the video features of the target area, the video features include the video features of each frame image in the video data, the differential features between two consecutive frames of images, and the cumulative differential features; based on the video features of the target area, the target is tracked to determine the perception data of the target area, the perception data includes the geometric data and trajectory data of the intruding target; based on the perception data of the target area, the three-dimensional space perception model of the power tower is updated, and the three-dimensional space perception model is visually displayed.
[0125] The embodiments described above 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 do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.
Claims
1. A control method for a three-dimensional space sensing tower electronic fence device, characterized in that: include: Monitor infrared images of the electronic fence area around power towers; Analyzing the infrared image to determine whether there is any illegal intrusion in the area around the power tower; If there is illegal intrusion, extract the video data of the target area where the illegal intrusion occurs, and analyze the video data frame by frame to determine the video features of the target area, wherein the video features include the video features of each frame image in the video data, the differential features between two consecutive frames of images, and the accumulated differential features; Based on the video features of the target area, target tracking is performed to determine the perception data of the target area, wherein the perception data includes geometric data and trajectory data of the intrusion target; Based on the perception data of the target area, the three-dimensional space perception model of the power tower is updated, and the three-dimensional space perception model is visually displayed.
2. The control method of the three-dimensional space perception tower electronic fence device according to claim 1 is characterized in that: The analyzing the infrared image to determine whether there is illegal intrusion in the area around the power tower includes: For any area in the area around the power tower, grayscale calculation is performed based on a standard image of the area to determine the grayscale threshold of each pixel in the area; the standard image is an image of the area when there is no intrusion; Converting the infrared image into a grayscale image, and calculating the grayscale value of each pixel in the grayscale image; Based on the grayscale value of each pixel in the grayscale image and the grayscale threshold of each pixel, threshold judgment is performed to determine abnormal pixels in the area; Based on the abnormal pixels in the area, determine the image block where the intrusion target is located; Based on the infrared image and the image block where the intrusion target is located, edge detection is performed to determine the geometric data and position information of the intrusion target; the geometric data includes the size and shape of the intrusion target, and the position information is used to characterize the relative position relationship between the center of the intrusion target and the electronic fence; Based on the geometric data and position information of the intrusion target, determine whether there is illegal intrusion in the area.
3. The control method of the three-dimensional space perception tower electronic fence device according to claim 1 is characterized in that: The step of analyzing the video data frame by frame to determine the video features of the target area includes: Performing frame processing on the video data to obtain multiple frames of video images; Extract features from each frame of video image to determine the global features and local features of each frame of video image; Calculating the pixel difference between two consecutive frames of video images in the multiple frames of video images to generate multiple frames of differential images; Performing feature extraction on the multi-frame differential images to determine the differential features, wherein the differential features include global differential features and local differential features; Accumulating the plurality of frames of differential images to obtain an accumulated differential image; Performing feature extraction on the cumulative difference image to determine the cumulative difference features of the cumulative difference image, wherein the cumulative difference features include global cumulative features and local cumulative features; The video features of the target area are generated based on the global features and local features of each frame of the video image, the differential features, and the accumulated differential features.
4. The control method of the three-dimensional space perception tower electronic fence device according to claim 3 is characterized in that: The step of extracting features from each frame of video image to determine global features and local features of each frame of video image includes: For any frame of video image, the global features of the frame of video image are calculated, and the global features include the histogram and color moment of the video image; the histogram is used to characterize the probability of occurrence of each gray level on the grayscale image of the video image; the color moment is used to characterize the distribution of each color on the video image; Traversing each pixel point of the frame of video image, taking each pixel point as the center and taking a first threshold number of pixels points as the radius, to determine the circumference corresponding to each pixel point; Compare the brightness of each pixel point on the circumference corresponding to each pixel point with the brightness of the central pixel point to determine the comparison result; If the comparison result satisfies a set condition, the central pixel point is determined as a candidate point, and the preset condition is that the difference between the brightness of a second threshold number of consecutive pixel points on the circumference and the brightness of the candidate point is greater than a set brightness difference; For multiple adjacent candidate points, non-maximum suppression is used to delete the candidate points with non-maximum values among the multiple adjacent candidate points, and the candidate points with maximum values are retained as key points to form a key point group; Calculate the grayscale centroid of the key point group based on the position of each key point in the key point group in the frame video image; Determine the main direction of each key point based on the grayscale centroid of the key point group and the position of each key point; Along the main direction of each key point, compare the brightness of each key point in the main direction to determine the flag of each key point; the flag is used to characterize the brightness change of the key point along the main direction; Based on the position, brightness, grayscale centroid and flag of each key point on the frame of video image, the local features of the frame of video image are determined.
5. The control method of the three-dimensional space perception tower electronic fence device according to claim 1 is characterized in that: The target tracking based on the video features of the target area and determining the perception data of the target area include: Inputting the video features of the target area into a preset target tracking model to obtain feature data of the intrusion target, wherein the feature data of the intrusion target includes geometric data of the intrusion target and position information of the intrusion target on each frame of the video image; Determine the actual position of the intrusion target in the target area based on the position information of the intrusion target on each frame of the video image; Based on the actual position of the intrusion target in the target area and the time corresponding to each frame of image, the trajectory data of the intrusion target is determined.
6. The control method of the three-dimensional space perception tower electronic fence device according to claim 5 is characterized in that: Before performing target tracking based on the video features of the target area and determining the perception data of the target area, the method further includes: Obtain video images of various intrusion objects in motion, as well as geometric data and location information of various intrusion objects; Based on the video images of various intruders in motion, frame-by-frame analysis is performed to generate video features of various intruders; The video features of various intrusions are used as input, and the geometric data and location information of various intrusions are used as output to generate training samples; A neural network training is performed based on the training samples to obtain a target tracking model.
7. The control method of the three-dimensional space perception tower electronic fence device according to claim 5 is characterized in that: The performing of neural network training based on the training samples to obtain a target tracking model includes: Performing neural network training based on the training samples to obtain an initial model; Taking the initial model as a teacher model, constructing a student model having the same model parameters as the teacher model; Testing the teacher model based on the test set in the training sample to obtain a first test accuracy of the teacher model; Based on the first test accuracy, determining a test accuracy threshold; Performing knowledge distillation on the student model to obtain a distilled student model; Testing the distilled student model based on the test set in the training sample to obtain a second test accuracy; If the second test accuracy is greater than or equal to the test accuracy threshold, repeatedly performing knowledge distillation on the student model until the second test accuracy is less than the test accuracy threshold; If the second test accuracy is less than the test accuracy threshold, the student model before the current student model knowledge distillation is used as the target tracking model.
8. The control method of the three-dimensional space perception tower electronic fence device according to claim 1 is characterized in that: The updating of the three-dimensional spatial perception model of the power tower based on the perception data of the target area includes: Based on the geometric data of the intrusion target in each frame of video image, three-dimensional modeling is performed to determine the three-dimensional model of the intrusion target; Based on the three-dimensional model of the intrusion target and the trajectory data of the intrusion target, the three-dimensional space perception model of the power tower is updated to obtain the three-dimensional space perception model at each moment.
9. The control method of the three-dimensional space perception tower electronic fence device according to claim 1, characterized in that: After performing target tracking based on the video features of the target area and determining the perception data of the target area, the method further includes: Based on the sensing data, intrusion judgment is performed to determine intrusion information, wherein the intrusion information includes type and quantity; Based on the intruder information, a database is queried to generate an expulsion control instruction; the expulsion control instruction is used to expel the intruder from the target area.
10. A three-dimensional space sensing tower electronic fence device, characterized in that: The electronic fence device includes an infrared detection device, a camera device and a control device, and the control device is used to execute the steps of the method as claimed in any one of claims 1 to 9.
