A power transmission equipment video detection system

Abstract

This invention provides a video detection system for power transmission equipment, including a camera acquisition module, a data analysis module, and a control module. The camera acquisition module acquires video monitoring data from the power transmission equipment and transmits the acquired video monitoring data to the data analysis module. The data analysis module performs image analysis processing based on the acquired video monitoring data, including anomaly feature identification based on the acquired regular monitoring data, obtaining anomaly feature identification results and generating feature detection control commands; and anomaly analysis based on the acquired feature video data, obtaining anomaly analysis results. The control module sends control commands to the corresponding camera acquisition modules according to the acquired feature detection control commands, so that the camera acquisition modules can increase the resolution of the current video monitoring data according to the control commands. This invention can reduce the consumption of acquisition tasks while ensuring the completion of anomaly detection tasks, thus optimizing the video detection system.

Description

Technical Field

[0001] This invention relates to the field of power detection technology, and in particular to a video detection system for power transmission equipment. Background Technology

[0002] Currently, most of the existing transmission lines have long routes, complex and varied terrain along the way, large regional spans, and long inspection times. This makes the transmission lines and related equipment susceptible to damage from external forces (such as accidental damage caused by nearby construction or environmental factors), leading to problems and failures in the transmission equipment and seriously affecting the safety of large-scale transmission line networks.

[0003] In existing technologies, anomaly detection of power transmission equipment (such as high-voltage towers and cables) mostly involves setting up video acquisition devices near the equipment to continuously collect video images and combine them with video analysis technology to complete the detection task. However, the continuous video data acquisition and transmission process consumes a significant amount of energy and data transmission resources. To ensure the detection effect, when a large number of image acquisition devices are deployed for the designed detection task, the overall energy and data transmission resource consumption cannot be ignored. Therefore, there is an urgent need to propose a technical solution that can optimize video detection technology for power transmission equipment. Summary of the Invention

[0004] To address the aforementioned problems, this invention aims to provide a video inspection system for power transmission equipment.

[0005] The objective of this invention is achieved through the following technical solution:

[0006] This invention discloses a video detection system for power transmission equipment, including a camera acquisition module, a data analysis module, and a control module;

[0007] The camera acquisition module is fixedly installed on the high-voltage tower and aimed at the designated power transmission equipment to acquire video monitoring data of the power transmission equipment and transmit the acquired video monitoring data to the data analysis module; the video monitoring data includes regular video data and feature video data, wherein the resolution of the regular video data is lower than that of the feature video data;

[0008] The data analysis module is used to perform image analysis processing based on the acquired video surveillance data, including identifying abnormal features based on the acquired regular monitoring data to obtain abnormal feature identification results; and performing anomaly analysis based on the acquired feature video data to obtain anomaly analysis results; when the abnormal feature identification result is abnormal, a feature detection control command is generated and transmitted to the control module; when the anomaly analysis result is abnormal, an anomaly analysis result for the power transmission equipment is generated and transmitted to the control center.

[0009] The control module is used to send control commands to the corresponding camera acquisition module according to the acquired feature detection control commands, so that the camera acquisition module can improve the resolution of the current video surveillance data according to the control commands.

[0010] Preferably, the camera acquisition module includes a camera unit and a communication unit;

[0011] The camera unit is fixedly mounted on the high-voltage tower and aimed at the designated power transmission equipment. It is used to acquire the conventional video data of the power transmission equipment using a preset first resolution, first frame rate and first bit rate; and to acquire the characteristic video data of the power transmission equipment by adjusting to a second resolution, second frame rate and second bit rate according to the received transmission control command.

[0012] The communication unit is used to transmit regular video data or feature video data acquired by the camera unit to the data analysis module through a wireless communication network.

[0013] Preferably, the camera unit further includes:

[0014] When a reset command is received or the duration of continuous acquisition of feature video data exceeds a set threshold, the system readjusts its resolution, frame rate, and bit rate, and acquires conventional video data of the power transmission equipment using a preset first resolution, first frame rate, and first bit rate.

[0015] Preferably, the data analysis module includes an anomaly identification unit, an anomaly analysis unit, a first output unit, and a second output unit;

[0016] The anomaly identification unit is used to identify anomaly features based on the acquired regular monitoring data, obtain the anomaly feature identification results, and transmit the anomaly feature identification results to the output unit;

[0017] The anomaly analysis unit is used to perform anomaly analysis based on the acquired feature video data, obtain the anomaly analysis results, and transmit the anomaly analysis results to the output unit.

[0018] The first output unit is used to generate a feature detection control command and transmit it to the control module when the abnormal feature identification result is abnormal. The abnormal feature identification result includes the identifier of the power transmission equipment with abnormal features; the feature detection control command includes the identifier of the corresponding camera acquisition module.

[0019] The second output unit is used to generate power transmission equipment anomaly analysis results and transmit them to the control center when anomaly analysis results are found.

[0020] Preferably, the anomaly detection unit includes:

[0021] Based on the acquired routine monitoring data, the regions are divided to obtain fixed and variable regions in the image;

[0022] For fixed regions, a preset region template is used to compare and analyze the fixed region image of the current image to obtain the fixed region anomaly identification result;

[0023] For indeterminate regions, anomaly identification is performed on the current image by comparing previous and next frames, and the anomaly identification results for indeterminate regions are obtained.

[0024] The anomaly identification results are obtained by combining the anomaly identification results of fixed regions and anomaly identification results of variable regions.

[0025] Preferably, the anomaly analysis unit includes:

[0026] Preprocess the acquired feature video data;

[0027] The preprocessed feature video data is divided into blocks, and the feature video frame is divided into multiple image blocks.

[0028] The obtained image patches are input into a preset anomaly analysis model. The anomaly analysis model detects anomalies based on the image patches, obtains anomaly analysis results, and marks the corresponding image patches.

[0029] Based on the anomaly analysis results of each image block in the video footage, the anomaly analysis results of the power transmission equipment are obtained by combining the results.

[0030] Preferably, the control module includes an instruction unit and an association unit;

[0031] The instruction unit is used to extract the camera acquisition module identifier based on the acquired feature detection control instruction, and send control instructions to the corresponding camera acquisition module according to the camera acquisition module identifier, so that the camera acquisition module can improve the resolution, frame rate and bit rate of the current video surveillance data according to the control instruction;

[0032] The association unit is used to extract the identifier of the power transmission equipment that exhibits abnormal features according to the feature detection control command, find the camera acquisition module associated with the power transmission equipment identifier according to the preset camera-power transmission equipment correspondence table, and send control commands to the associated camera acquisition module so that the camera acquisition module can improve the resolution, frame rate and bit rate of the current video monitoring data according to the control command.

[0033] Preferably, the system also includes a control center, wherein the control center includes a display unit;

[0034] The display unit is used to showcase the obtained anomaly analysis results of power transmission equipment.

[0035] Preferably, the system also includes an environmental monitoring module;

[0036] The environmental monitoring module is installed around the high-voltage tower to collect environmental monitoring data of the power transmission equipment and transmit the collected environmental monitoring data to the data analysis module.

[0037] The data analysis module is also used to perform environmental anomaly analysis based on the acquired environmental monitoring data and obtain environmental anomaly analysis results; when an anomaly is found in the environmental anomaly analysis results, a feature detection control command is generated and transmitted to the control module.

[0038] Preferably, the environmental monitoring module includes a sensor unit;

[0039] The sensor unit includes various types of sensors installed around the high-voltage tower. These sensors collect environmental monitoring data about the environment where the power transmission equipment is located. The sensors include microwave sensors, smoke sensors, vibration sensors, acceleration sensors, and induction alarm sensors. The environmental monitoring data includes microwave detection data, smoke detection data, vibration detection data, acceleration detection data, and induction detection data.

[0040] The data analysis module also includes an environmental analysis unit;

[0041] The environmental analysis unit is used to compare and analyze the acquired environmental monitoring data with preset environmental indicators. When one or more environmental monitoring data exceed the preset environmental indicator range, the environmental anomaly analysis result is abnormal; otherwise, the environmental anomaly analysis result is normal.

[0042] The beneficial effects of this invention are as follows: This invention proposes a video inspection system for power transmission equipment. During routine inspection, the camera acquisition module can be configured to acquire video inspection data in a low-power mode with lower resolution, frame rate, and bit rate, effectively reducing the consumption of image data acquired by the camera during normal / normal time. The data analysis module performs adaptive anomaly feature recognition based on the obtained routine monitoring data, enabling the identification of abnormal targets in the image based on low-resolution image data. When an abnormal target is identified, the control module further controls the corresponding camera acquisition module to increase the resolution, frame rate, and bit rate to acquire clearer and more coherent video inspection data. Further analysis of power transmission equipment anomalies based on the acquired feature video data effectively ensures the accuracy and reliability of anomaly analysis based on video images. Simultaneously, the camera acquisition module mode switching method reduces the consumption of video monitoring data acquisition tasks under normal conditions while ensuring the completion of anomaly detection tasks, thus optimizing the video inspection system. This system meets the needs of power transmission equipment inspection in large-scale power transmission network. Attached Figure Description

[0043] The present invention will be further described with reference to the accompanying drawings, but the embodiments in the drawings do not constitute any limitation on the present invention. For those skilled in the art, other drawings can be obtained based on the following drawings without creative effort.

[0044] Figure 1 This is a structural diagram of a video detection system for power transmission equipment according to an embodiment of the present invention;

[0045] Figure 2 This is a structural diagram of a video detection system for power transmission equipment according to another embodiment of the present invention. Detailed Implementation

[0046] The present invention will be further described in conjunction with the following application scenarios.

[0047] See Figure 1 The embodiment illustrates a video detection system for power transmission equipment, including a camera acquisition module, a data analysis module, and a control module;

[0048] The camera acquisition module is fixedly installed on the high-voltage tower and aimed at the designated power transmission equipment to acquire video monitoring data of the power transmission equipment and transmit the acquired video monitoring data to the data analysis module. The video monitoring data includes regular video data and feature video data, wherein the resolution, frame rate and bit rate of the regular video data are lower than those of the feature video data.

[0049] The data analysis module is used to perform image analysis processing based on the acquired video surveillance data, including identifying abnormal features based on the acquired regular monitoring data to obtain abnormal feature identification results; and performing anomaly analysis based on the acquired feature video data to obtain anomaly analysis results; when the abnormal feature identification result is abnormal, a feature detection control command is generated and transmitted to the control module; when the anomaly analysis result is abnormal, an anomaly analysis result for the power transmission equipment is generated and transmitted to the control center.

[0050] The control module is used to send control commands to the corresponding camera acquisition module according to the acquired feature detection control commands, so that the camera acquisition module can improve the resolution of the current video surveillance data according to the control commands.

[0051] In the above embodiments of the present invention, during routine inspection, the camera acquisition module can be configured to acquire video detection data in a low-power mode with lower resolution, frame rate, and bit rate, effectively reducing the consumption of image data acquired by the camera during normal / normal time. The data analysis module performs adaptive anomaly feature identification based on the obtained routine monitoring data, enabling the identification of abnormal targets in the image based on low-resolution image data. When an abnormal target is identified, the control module further controls the corresponding camera acquisition module to increase the resolution, frame rate, and bit rate to acquire clearer and more coherent video detection data. Further analysis of the acquired feature video data on the power transmission equipment effectively ensures the accuracy and reliability of power transmission equipment anomaly analysis based on video images. Simultaneously, the camera acquisition module mode switching method reduces the consumption of video monitoring data acquisition tasks under normal conditions while ensuring the completion of anomaly detection tasks, thus optimizing the video detection system. This meets the needs of power transmission equipment inspection in large-scale power transmission network.

[0052] Different types of video surveillance data carry corresponding video data type tags (regular video data, feature video data). The data analysis module performs image analysis processing on the received video surveillance data in a way that corresponds to the type tag.

[0053] Preferably, the system also includes a control center, wherein the control center includes a display unit;

[0054] The display unit is used to showcase the obtained anomaly analysis results of power transmission equipment.

[0055] Through the control center, managers can obtain real-time anomaly analysis results for power transmission equipment transmitted by the video detection system, and make corresponding scheduling and maintenance task assignments based on the anomaly analysis results, so as to further handle the abnormal situations of power transmission equipment in a reasonable manner.

[0056] Preferably, the camera acquisition module includes a camera unit and a communication unit;

[0057] The camera unit is fixedly mounted on the high-voltage tower and aimed at the designated power transmission equipment. It is used to acquire the conventional video data of the power transmission equipment using a preset first resolution, first frame rate and first bit rate; and to acquire the characteristic video data of the power transmission equipment by adjusting to a second resolution, second frame rate and second bit rate according to the received transmission control command.

[0058] The communication unit is used to transmit regular video data or feature video data acquired by the camera unit to the data analysis module through a wireless communication network.

[0059] The camera unit is selected from smart cameras capable of switching between different working modes. Depending on the actual situation, the camera unit can be set to complete the acquisition and transmission of regular video data in a low-power mode with a resolution of 640×480, a frame rate of 5fps, and a bit rate of 1024Kbps, and to complete the acquisition and transmission of feature video data in a fine mode with a resolution of 2560×1440, a frame rate of 20fps, and a bit rate of 12288Kbps.

[0060] It should be noted that the difference between regular video data and feature video data collected by the same camera unit lies only in the different settings such as resolution, frame rate, and bit rate. The specific area covered and the shooting angle of the camera unit remain unchanged.

[0061] When the camera unit operates in low-power mode, it can greatly reduce the energy consumed and data transmission volume required for the camera unit to collect video surveillance data, thus effectively reducing the energy consumption of the camera unit.

[0062] Preferably, the camera unit further includes:

[0063] When a reset command is received or the duration of continuous acquisition of feature video data exceeds a set threshold, the system readjusts its resolution, frame rate, and bit rate, and acquires conventional video data of the power transmission equipment using a preset first resolution, first frame rate, and first bit rate.

[0064] After receiving an instruction or after a set time interval has elapsed, the camera unit can automatically return to low-power mode to collect video surveillance data, thereby ensuring the optimization of power consumption for video surveillance data collection.

[0065] Preferably, the data analysis module includes an anomaly identification unit, an anomaly analysis unit, a first output unit, and a second output unit;

[0066] The anomaly identification unit is used to identify anomaly features based on the acquired regular monitoring data, obtain the anomaly feature identification results, and transmit the anomaly feature identification results to the output unit;

[0067] The anomaly analysis unit is used to perform anomaly analysis based on the acquired feature video data, obtain the anomaly analysis results, and transmit the anomaly analysis results to the output unit.

[0068] The first output unit is used to generate a feature detection control command and transmit it to the control module when the abnormal feature identification result is abnormal. The abnormal feature identification result includes the identifier of the power transmission equipment with abnormal features; the feature detection control command includes the identifier of the corresponding camera acquisition module.

[0069] The second output unit is used to generate power transmission equipment anomaly analysis results and transmit them to the control center when anomaly analysis results are found.

[0070] The data analysis module employs an anomaly identification unit and an anomaly analysis unit to process video surveillance data acquired under two different modes. For low-resolution conventional monitoring data, anomaly feature identification technology is used to detect suspected abnormal targets in the image. When a suspected abnormal target is detected, the first output unit sends a feature detection control command to the camera acquisition module to acquire clearer feature image data for further anomaly analysis. The anomaly analysis unit, based on the feature image data and an anomaly analysis model built using deep learning and other technologies, analyzes the obtained feature image data to accurately identify anomalies in the image. The second output unit transmits the anomaly analysis results of the power transmission equipment to the control center for further warnings and alerts. By combining "coarse" identification of suspected abnormal targets based on low-resolution images with "fine" analysis of anomalies based on high-resolution images, the reliability and accuracy of anomaly analysis of power transmission equipment based on video detection can be ensured while effectively reducing the energy consumption of data processing and video monitoring data acquisition, thus contributing to energy consumption optimization of the power transmission equipment video detection system. This approach meets the needs of power transmission equipment detection in large-scale power transmission network.

[0071] Preferred, see Figure 2 As shown in the embodiment, the system also includes an environmental monitoring module;

[0072] The environmental monitoring module is installed around the high-voltage tower to collect environmental monitoring data of the power transmission equipment and transmit the collected environmental monitoring data to the data analysis module.

[0073] The data analysis module is also used to perform environmental anomaly analysis based on the acquired environmental monitoring data and obtain environmental anomaly analysis results; when an anomaly is found in the environmental anomaly analysis results, a feature detection control command is generated and transmitted to the control module.

[0074] Based on video detection technology for power transmission equipment inspection, the system can also collaborate with an environmental monitoring module. This module acquires environmental monitoring data of the power transmission equipment, and the system uses this data to conduct preliminary analysis of any anomalies. When the environmental monitoring data indicates potential anomalies affecting the equipment, the system further controls the camera acquisition module to collect detailed video data and performs further anomaly analysis based on the characteristic video data. By using environmental monitoring data as a control factor for the camera acquisition module, the video detection system can be further optimized, improving its adaptability (e.g., by integrating with existing conventional online power transmission line inspection systems).

[0075] Preferably, the environmental monitoring module includes a sensor unit;

[0076] The sensor unit includes various types of sensors installed around the high-voltage tower. These sensors collect environmental monitoring data of the environment where the power transmission equipment is located. The sensors include microwave sensors, smoke sensors, vibration sensors, acceleration sensors, and induction alarm sensors. The environmental monitoring data includes microwave detection data, smoke detection data, vibration detection data, acceleration detection data, and induction detection data.

[0077] Preferably, the data analysis module also includes an environmental analysis unit;

[0078] The environmental analysis unit is used to compare and analyze the acquired environmental monitoring data with preset environmental indicators. When one or more environmental monitoring data exceed the preset environmental indicator range, the environmental anomaly analysis result is abnormal; otherwise, the environmental anomaly analysis result is normal.

[0079] Based on the acquired environmental monitoring data, the data is compared and analyzed with relevant indicators. When anomalies are detected, corresponding control commands are generated to instruct the corresponding camera acquisition modules to increase resolution and further acquire characteristic video data of the relevant power transmission equipment. Based on this characteristic video data, a final situation analysis is performed, improving the reliability of power transmission equipment anomaly detection. This also enhances the adaptability and intelligence of the camera acquisition module control.

[0080] Preferably, the anomaly detection unit includes:

[0081] Based on the acquired routine monitoring data, the regions are divided to obtain fixed and variable regions in the image;

[0082] For fixed regions, a preset region template is used to compare and analyze the fixed region image of the current image to obtain the fixed region anomaly identification result;

[0083] For indeterminate regions, anomaly identification is performed on the current image by comparing previous and next frames, and the anomaly identification results for indeterminate regions are obtained.

[0084] The anomaly identification results are obtained by combining the anomaly identification results of fixed regions and anomaly identification results of variable regions.

[0085] In the preparation phase, given the relatively fixed position of power transmission equipment in the video surveillance frame, area labeling is performed based on video surveillance data samples acquired by the camera acquisition module. Areas involving power transmission equipment are marked as fixed areas, while the remaining background areas are marked as variable areas. This facilitates rapid segmentation of the power transmission equipment and background areas under local resolution conditions, and allows for separate anomaly identification processing based on the different areas. This adapts to the needs of anomaly identification in conventional video data acquired from different shooting angles and with different power transmission equipment. For the fixed areas, i.e., the areas where the power transmission equipment is located, template matching can quickly identify whether there are abnormal coverings on the equipment. For the variable areas, anomalies are identified by comparing consecutive frames. When anomalies are detected, the camera acquisition module can be further controlled to adjust the resolution and acquire further feature video data for further anomaly analysis and processing. This helps reduce the pressure on monitoring data acquisition and transmission, and improves system performance.

[0086] Preferably, in the anomaly identification unit, the area is divided based on the acquired regular monitoring data, specifically including:

[0087] The acquired regular monitoring data is divided into regions based on a preset division template, resulting in fixed regions and variable regions.

[0088] The partitioning template is used during the setup and debugging phase. Based on the video surveillance data acquired by the camera acquisition module, the area of ​​the power transmission equipment in the video surveillance data is marked as the equipment area, and the remaining area is marked as the background area.

[0089] When dividing the acquired regular monitoring data into regions, the division template corresponding to the same camera acquisition module is applied to the current regular monitoring data screen. The part of the fixed area in the corresponding template is marked as the fixed area, and the part of the background area in the corresponding template is marked as the variable area.

[0090] Considering that the camera unit captures video surveillance data from a fixed angle after setup, and the power transmission equipment is positioned relatively fixed (i.e., its position in the video surveillance data is fixed), during the setup and debugging phase, the area containing the power transmission equipment can be marked as the foreground area based on the video surveillance data captured by the camera unit, and the remaining areas can be marked as the background area. This establishes a segmentation template, which facilitates subsequent segmentation of the video surveillance data into regions, improving the efficiency and accuracy of region segmentation.

[0091] Meanwhile, when dividing the area, the target image is further extracted from the power transmission equipment area as the standard area template for the fixed area, so that it can be called in subsequent anomaly identification based on the fixed area.

[0092] Preferably, in the anomaly detection unit, a pre-set region template is compared and analyzed with a fixed region image of the current image, specifically including:

[0093] Based on the defined fixed regions, retrieve the region template for each region, and compare and analyze the retrieved region template with the fixed region image of the current image.

[0094] Local similarity analysis is performed on the region template and the obtained fixed region image to obtain the local similarity between the fixed region image and the region template. When the local similarity is greater than the preset standard, the fixed region anomaly identification result is normal; otherwise, when there is a local similarity less than the preset standard, the fixed region anomaly identification result is abnormal.

[0095] Local similarity analysis includes:

[0096] Map the pixels in the fixed region image A and the region template B to a unified coordinate system;

[0097] The grayscale value and RGB color channel value of each pixel are obtained respectively, and the similarity feature factor of each pixel is calculated. The similarity feature factor calculation function used is:

[0098]

[0099] in, Let h represent the similarity feature factor of pixel (x, y). In (a, b) ∈ bis(x, y), bis(x, y) represents the set of pixels within a 3×3 region centered at pixel (x, y), and pixel (a, b) is a pixel belonging to the set bis(x, y). A (x, y) represents the gray value of pixel (x, y) in a fixed region of the image, h B(a, b) represents the grayscale value at pixel position (a, b) in region template B, r A (x, y), g A (x, y), b A (x, y) represent the r, g, and b channel values ​​of the pixel at position (x, y) in a fixed region of the image, respectively. B (a, b), g B (a, b), b B (a, b) represent the r, g, and b channel values ​​at the pixel position (a, b) in region template B, respectively;

[0100] A feature window Q with a window size of n×n is used to traverse a fixed region image sequentially to obtain the local similarity of each local location. The traversal method includes aligning the center of the feature window with each pixel in the fixed region image sequentially to complete the traversal, or moving the feature window with a preset fixed step size to obtain the local similarity of the corresponding location.

[0101] The local similarity feature calculation function used for the feature window is as follows:

[0102]

[0103] In the formula, y(c, d) represents the local similarity when the center of the feature window is aligned with the pixel (c, d), and mean(h) = 1 / 2. Q The mean(h) represents the average similarity feature factor of each pixel position in the feature window. A ) and mean(h B ) represent the average grayscale value of each pixel position in the fixed region image and the region template, respectively. represents the standard deviation of the similarity feature factor at each pixel position in the feature window, n represents the size of the feature window; kt(Q) represents the complexity correction coefficient, where sth(Q) and sts(Q) represent the average horizontal and vertical gray-level gradient differences of each pixel within the feature window, respectively; stT represents the preset standard value of gray-level gradient difference; β represents the correction parameter, where β∈[2,4];

[0104] The obtained local similarities are compared with the preset similarity standard value. When all local similarities are greater than the preset similarity threshold, the fixed region anomaly identification result is normal. Otherwise, when there is a local similarity less than the preset similarity threshold, the fixed region anomaly identification result is abnormal.

[0105] Preferably, the horizontal gray-level gradient difference and the vertical gray-level gradient difference are respectively determined by the gradient operator. and The result of the calculation is that

[0106] Preferably, the size of the feature window is 3×3, 4×4, 5×5, etc.

[0107] The present invention proposes a technical solution for anomaly detection in the area of ​​power transmission equipment in video surveillance data based on low-resolution images by performing region template comparison analysis on fixed areas divided from conventional monitoring data. This solution analyzes the local similarity between a fixed region image and its corresponding region template to identify anomalies in the power transmission equipment area within the video surveillance data. Specifically, considering that traditional similarity analysis on low-resolution images typically uses pixel grayscale comparison, which can lead to insufficient accuracy due to large image grains, the proposed local similarity analysis solution calculates the similarity feature factor of each pixel after aligning the coordinates of the fixed region image and the region template. A similarity feature factor calculation function is proposed, which performs a small-range local search based on the grayscale and RGB color channel features of the pixels, first filtering out matching images. Pixel combination can effectively improve the robustness of local similarity comparison, avoiding the problem of pixel-level grayscale value deviation caused by traditional large-grain images, thus improving reliability and accuracy. By extracting similarity features based on the grayscale characteristics and RGB color space characteristics of pixel combination, and by adding RGB features as a reference parameter, it can avoid "false grayscale similarity" (for complex images, matching based on a single grayscale parameter can easily result in the same grayscale value but different actual image content that cannot be detected), thus helping to improve the reliability of the similarity feature detection process. Based on the acquired similarity feature factors, a feature window is further used to calculate local similarity features of the image. The proposed local similarity feature calculation function accurately reflects the regions with abnormal objects by leveraging the local characteristics of the similarity feature factors. It can also adapt to similarity deviations caused by differences in illumination between the actual image and the template image. Considering that the illumination deviation phenomenon is prone to misjudgment in parts with many details in the power transmission equipment area (because the grayscale changes in the details affect the representation level of the illumination deviation), a complexity correction coefficient is specially added to improve the similarity comparison of parts with many detailed features. This further improves the adaptability and reliability of local similarity comparison and enhances the accuracy of abnormal object identification in the power transmission equipment area.

[0108] Preferably, in the anomaly recognition unit, for indefinite regions, anomaly recognition is performed on the current image by comparing previous and subsequent frames, specifically including:

[0109] Image frames are extracted from the acquired routine monitoring data. Based on the preceding and following image frames, and using the frame difference method, moving targets in the images are detected and marked.

[0110] The size of the moving target is extracted from the image region of the detected moving target. If the size of the moving target is larger than the preset standard size, the moving target is marked as an anomaly, and the anomaly identification result of the indefinite region is output as an anomaly.

[0111] By detecting moving targets, the background area near power transmission equipment can be monitored simultaneously. When an abnormal target that may pose a danger appears in the vicinity of the power transmission line, a more detailed analysis can be conducted based on the anomaly. This can effectively identify floating or flying objects near the power transmission line.

[0112] When both the fixed region anomaly identification result and the variable region anomaly identification result are normal, the output anomaly feature identification result is normal; otherwise, when either the fixed region anomaly identification result or the variable region anomaly identification result is abnormal, the output anomaly feature identification result is abnormal.

[0113] Preferably, the anomaly analysis unit includes:

[0114] Preprocess the acquired feature video data;

[0115] The preprocessed feature video data is divided into blocks, and the feature video frame is divided into multiple image blocks.

[0116] The obtained image patches are input into a preset anomaly analysis model. The anomaly analysis model detects anomalies based on the image patches, obtains anomaly analysis results, and marks the corresponding image patches.

[0117] Based on the anomaly analysis results of each image block in the video footage, the anomaly analysis results of the power transmission equipment are obtained by combining the results.

[0118] For high-definition feature video data, the images are first divided into blocks. Based on these blocks, a deep learning-trained anomaly recognition model is used to identify anomalies in each block. By dividing the images into blocks, the anomalies can be identified quickly and accurately, adapting to the characteristics of different anomalies and improving the effectiveness of anomaly analysis.

[0119] Preferably, in the anomaly analysis unit, preprocessing is performed based on the acquired feature video data, including:

[0120] Image processing techniques are used to enhance the feature video data, thereby improving its clarity.

[0121] Enhancing feature video data improves image clarity, helps eliminate interference during image acquisition or transmission, and improves image quality.

[0122] Preferably, in the anomaly analysis unit, the preprocessed feature video data is divided into blocks for processing, specifically including:

[0123] Based on the preset image size, the image is divided into N image blocks that have overlapping parts.

[0124] The obtained image patches are input into a preset anomaly analysis model, which then performs anomaly detection based on the image patches.

[0125] The anomaly analysis model is built on a neural network. It uses deep learning to train the model using an image training set during the training phase and tests the trained model using a test set. When the model's accuracy in identifying anomalies in images is higher than a preset standard, the anomaly analysis model is obtained. The image training set contains images of various possible anomalies (such as birds, drones, coverings, leaves, fires, etc.) and corresponding anomaly labels.

[0126] When an anomaly is detected based on any of the image blocks, the anomaly analysis results of the power transmission equipment are output according to the type of anomaly.

[0127] Optionally, the anomaly analysis model can also use an image recognition model that has been trained in the prior art to detect, identify and label anomalies or abnormal objects. This invention does not make any specific limitations here.

[0128] Preferably, adjacent image blocks on the left and right have 50% regional overlap (i.e., the right half of the left image block and the left half of the right image block overlap), and adjacent image blocks on the top and bottom have 50% regional overlap.

[0129] Preferably, the control module includes an instruction unit;

[0130] The instruction unit is used to extract the camera acquisition module identifier based on the acquired feature detection control instruction, and send control instructions to the corresponding camera acquisition module according to the camera acquisition module identifier, so that the camera acquisition module can improve the resolution, frame rate and bit rate of the current video surveillance data according to the control instruction.

[0131] The control module sends control commands to the corresponding camera acquisition module based on the received feature detection control commands, so as to control the camera module to improve the quality of image data acquisition, so as to further complete the anomaly analysis of the power transmission equipment based on the high-quality feature video data.

[0132] Preferably, the control module further includes an association unit;

[0133] The association unit is used to extract the identifier of the power transmission equipment that exhibits abnormal features according to the feature detection control command, find the camera acquisition module associated with the power transmission equipment identifier according to the preset camera-power transmission equipment correspondence table, and send control commands to the associated camera acquisition module so that the camera acquisition module can improve the resolution, frame rate and bit rate of the current video monitoring data according to the control command.

[0134] In another scenario, based on the obtained feature detection control instructions, for power transmission equipment that may identify abnormal targets, all camera acquisition modules associated with the power transmission equipment (e.g., all camera acquisition modules that acquire data from the power transmission equipment from different angles) are further designated as target devices. At the same time, the video surveillance image acquisition quality of all target devices is improved, and further anomaly analysis is performed based on the feature video data of the power transmission equipment to accurately detect the occurrence of abnormal situations in the power transmission equipment.

[0135] For example, when a possible abnormal covering is detected in power transmission equipment A based on the regular video data acquired by a certain camera acquisition module, the system retrieves information from all associated camera acquisition modules of power transmission equipment A (for example, by pre-associating all camera acquisition modules that appear in the video surveillance footage of power transmission equipment A with power transmission equipment A, or by associating the information of the power transmission equipment appearing in the video surveillance footage acquired by the camera acquisition module with the camera acquisition module itself), and further controls all associated camera acquisition modules to improve the quality of video surveillance data, and further performs anomaly analysis and processing on the abnormal covering situation of power transmission equipment A based on the feature video data.

[0136] It should be noted that the functional units / modules in the various embodiments of the present invention can be integrated into one processing unit / module, or each unit / module can exist physically separately, or two or more units / modules can be integrated into one unit / module. The integrated unit / module described above can be implemented in hardware or in the form of software functional units / modules.

[0137] From the above description of the embodiments, those skilled in the art will clearly understand that the embodiments described herein can be implemented in hardware, software, firmware, middleware, code, or any suitable combination thereof. For hardware implementation, the processor can be implemented in one or more of the following units: Application-Specific Integrated Circuit (ASIC), Digital Signal Processor (DSP), Digital Signal Processing Device (DSPD), Programmable Logic Device (PLD), Field-Programmable Gate Array (FPGA), processor, controller, microcontroller, microprocessor, other electronic units designed to implement the functions described herein, or combinations thereof. For software implementation, some or all of the processes of the embodiments can be implemented by a computer program instructing the associated hardware. During implementation, the program can be stored in a computer-readable medium or transmitted as one or more instructions or code on a computer-readable medium. Computer-readable media include computer storage media and communication media, wherein communication media include any medium that facilitates the transmission of a computer program from one place to another. Storage media can be any available medium accessible to a computer. Computer-readable media can include, but is not limited to, RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code having the form of instructions or data structures and accessible to a computer.

[0138] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit the scope of protection of the present invention. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should be able to analyze that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the essence and scope of the technical solutions of the present invention.

Claims

1. A video inspection system for power transmission equipment, characterized in that, Includes a camera acquisition module, a data analysis module, and a control module; The camera acquisition module is fixedly installed on the high-voltage tower and aimed at the designated power transmission equipment to acquire video monitoring data of the power transmission equipment and transmit the acquired video monitoring data to the data analysis module. The video surveillance data includes regular video data and feature video data, with the resolution of regular video data being lower than that of feature video data. The data analysis module is used to perform image analysis and processing based on the acquired video surveillance data, including anomaly feature identification based on the acquired regular surveillance data, and to obtain anomaly feature identification results; And perform anomaly analysis based on the acquired feature video data to obtain anomaly analysis results; when the anomaly feature recognition result is abnormal, generate feature detection control instructions and transmit them to the control module; When an anomaly analysis result is found, a power transmission equipment anomaly analysis result is generated and transmitted to the control center; including: A pre-set region template is used to compare and analyze a fixed region of the current image. Fixed region image and area template The pixels in the graph are mapped to a unified coordinate system; Obtain the grayscale value and RGB color channel value of each pixel, and calculate the similarity feature factor of each pixel: in, Represents pixels Similarity feature factors middle, Represented by pixels Centered The set of pixels within a region, pixels Belonging to set Pixels within; Represents pixels in a fixed region of an image. The grayscale value of the location, Represents the pixels in region template B The grayscale value of the location, , , These represent pixels in a fixed region of the image. Location Channel value, , , These represent the pixels in region template B. Location Channel value; Use window size Feature window By iterating through the fixed region of the image, the local similarity of each local location is obtained: In the formula, This indicates that the center of the feature window is aligned with the pixel. Local similarity corresponding to location. This represents the average similarity feature factor of each pixel position in the feature window. and These represent the average grayscale values ​​of each pixel position in the fixed region image and the region template, respectively. The standard deviation of the similarity feature factor at each pixel position in the feature window. Indicates the size of the feature window; Represents the complexity correction factor, where , and These represent the average horizontal and average vertical gray-level gradient differences at each pixel location within the feature window, respectively; where... This represents the preset standard value for grayscale gradient difference; Indicates the correction parameter, where ; The obtained local similarities are compared with the preset similarity standard value. When all local similarities are greater than the preset similarity threshold, the fixed region anomaly identification result is normal. Otherwise, when there is a local similarity less than the preset similarity threshold, the fixed region anomaly identification result is abnormal. The control module is used to send control commands to the corresponding camera acquisition module according to the acquired feature detection control commands, so that the camera acquisition module can improve the resolution of the current video surveillance data according to the control commands.

2. The video detection system for power transmission equipment according to claim 1, characterized in that, The camera acquisition module includes a camera unit and a communication unit; The camera unit is fixedly installed on the high-voltage tower and aimed at the designated power transmission equipment, and is used to collect the conventional video data of the power transmission equipment using a preset first resolution, first frame rate and first bit rate. And for acquiring characteristic video data of power transmission equipment by adjusting to a second resolution, a second frame rate, and a second bit rate according to the received transmission control command; The communication unit is used to transmit regular video data or feature video data acquired by the camera unit to the data analysis module through a wireless communication network.

3. The video detection system for power transmission equipment according to claim 2, characterized in that, The camera unit also includes: When a reset command is received or the duration of continuous acquisition of feature video data exceeds a set threshold, the system readjusts its resolution, frame rate, and bit rate, and acquires conventional video data of the power transmission equipment using a preset first resolution, first frame rate, and first bit rate.

4. The video detection system for power transmission equipment according to claim 1, characterized in that, The first output unit is used to generate a feature detection control command and transmit it to the control module when the abnormal feature identification result is abnormal. The abnormal feature identification result includes the identifier of the power transmission equipment with abnormal features; the feature detection control command includes the identifier of the corresponding camera acquisition module. The second output unit is used to generate power transmission equipment anomaly analysis results and transmit them to the control center when anomaly analysis results are found.

5. The video detection system for power transmission equipment according to claim 1, characterized in that, The anomaly analysis unit includes: Preprocess the acquired feature video data; The preprocessed feature video data is divided into blocks, dividing the feature video frame into multiple image blocks. The obtained image patches are input into a preset anomaly analysis model. The anomaly analysis model detects anomalies based on the image patches, obtains anomaly analysis results, and marks the corresponding image patches. Based on the anomaly analysis results of each image block in the video footage, the anomaly analysis results of the power transmission equipment are obtained by combining the results.

6. The video detection system for power transmission equipment according to claim 1, characterized in that, The control module includes an instruction unit and an association unit; The instruction unit is used to extract the camera acquisition module identifier based on the acquired feature detection control instruction, and send control instructions to the corresponding camera acquisition module according to the camera acquisition module identifier, so that the camera acquisition module can improve the resolution, frame rate and bit rate of the current video surveillance data according to the control instruction; The association unit is used to extract the identifier of the power transmission equipment that exhibits abnormal features according to the feature detection control command, find the camera acquisition module associated with the power transmission equipment identifier according to the preset camera-power transmission equipment correspondence table, and send control commands to the associated camera acquisition module so that the camera acquisition module can improve the resolution, frame rate and bit rate of the current video monitoring data according to the control command.

7. The video detection system for power transmission equipment according to claim 1, characterized in that, It also includes a control center, which includes a display unit; The display unit is used to showcase the acquired results of power transmission equipment anomaly analysis.

8. The video detection system for power transmission equipment according to claim 1, characterized in that, It also includes an environmental monitoring module; The environmental monitoring module is installed around the high-voltage tower to collect environmental monitoring data of the power transmission equipment and transmit the collected environmental monitoring data to the data analysis module. The data analysis module is also used to perform environmental anomaly analysis based on the acquired environmental monitoring data and obtain the environmental anomaly analysis results; When an anomaly is detected in the environmental anomaly analysis results, a feature detection control command is generated and transmitted to the control module.

9. A video detection system for power transmission equipment according to claim 8, characterized in that, The environmental monitoring module includes sensor units; The sensor unit includes various types of sensors installed around the high-voltage tower. These sensors collect environmental monitoring data of the environment where the power transmission equipment is located. The sensors include microwave sensors, smoke sensors, vibration sensors, acceleration sensors, and induction alarm sensors. The environmental monitoring data includes microwave detection data, smoke detection data, vibration detection data, acceleration detection data, and induction detection data. The data analysis module also includes an environmental analysis unit; The environmental analysis unit is used to compare and analyze the acquired environmental monitoring data with preset environmental indicators. When one or more environmental monitoring data exceed the preset environmental indicator range, the environmental anomaly analysis result is abnormal; otherwise, the environmental anomaly analysis result is normal.

Images