Power transmission line inspection method and equipment

Abstract

The invention provides an inspection method for a power transmission line. The inspection method comprises the following steps: acquiring a first round of acquisition data set and carrying out optimization processing to generate a first round of initial data set; identifying abnormal images in the first round of initial data set to form an abnormal image data set; according to the data content of the abnormal image data set, determining the type of an environmental factor causing the abnormal image; according to the type of the environmental factor causing the abnormal image, formulating a corresponding sensing data acquisition task, and assigning the task to each sensor to execute a second round of sensing data acquisition; performing fusion, optimization and safety processing on the sensor data acquired in the second round for transmission; and decrypting the received data to obtain inspection data of the power transmission line equipment. The power transmission line inspection method and equipment provided by the invention can better adapt to complex terrains and changeable climate conditions, more accurately reflect the actual operation state of the power transmission line, improve the inspection efficiency and find equipment faults in time.

Description

Technical Field

[0001] This application relates to the field of power system testing, and in particular to a method and equipment for inspecting transmission lines. Background Technology

[0002] Transmission lines are a crucial component of the power system, and their operational status directly impacts the safety, stability, and reliability of the power grid. Therefore, regular inspections of transmission lines are of paramount importance. Current transmission line inspection methods primarily rely on manual inspections or single types of sensor equipment. However, these methods are poorly adaptable to complex terrain and variable weather conditions. For example, in scenarios with poor visible light imaging conditions, such as fog, strong sunlight, or nighttime, visible light cameras struggle to acquire clear images; infrared or ultraviolet equipment may also experience data distortion due to environmental noise interference. These factors make inspection results susceptible to external influences, making it difficult to accurately reflect the actual operating status of the transmission lines, thus affecting inspection efficiency and the timeliness of equipment fault detection. Summary of the Invention

[0003] This application proposes a method and equipment for inspecting power transmission lines, which can solve the problems existing in the background art.

[0004] To achieve the above objectives, this application adopts the following technical solution:

[0005] Firstly, a method for inspecting power transmission lines is provided, comprising the following steps:

[0006] S1. Obtain the first round of data collection dataset of the transmission line obtained in the first round of data collection, including images of the transmission line and environmental sensor data corresponding to the images;

[0007] S2. Optimize the first round of collected datasets to generate the first round of initial datasets;

[0008] S3. Identify anomalous images in the initial dataset of the first round to form an anomalous image dataset;

[0009] S4. Based on the data content of the abnormal image dataset, determine the type of environmental factors that caused the abnormal images;

[0010] S5. Based on the type of environmental factors that cause abnormal images, formulate corresponding sensor data acquisition tasks and assign these tasks to each sensor to perform the second round of sensor data acquisition.

[0011] S6. Acquire the sensor data obtained in the second round of acquisition, fuse and process it to obtain the initial dataset for the second round;

[0012] S7. Optimize and securely process the initial dataset for the second round, generate a secure data packet for the second round, and complete the transmission.

[0013] S8. Decrypt the received second-round security data packet to obtain the corresponding inspection data of the transmission line equipment.

[0014] Step S2, which involves optimizing the first round of collected datasets to generate the first round of initial datasets, includes the following steps:

[0015] A1. A weighted average fusion algorithm is used to perform pixel-level fusion of infrared and visible light images in the images of transmission lines to form a fused image set;

[0016] A2. Using histogram equalization, the images in the fused image set are enhanced by adjusting the image contrast to obtain the enhanced image set;

[0017] A3. Compare the pixel value variance of the images in the enhanced image set with a first preset threshold, filter out images whose pixel value variance is lower than the first preset threshold, and use median filtering to denoise them to obtain a net image set;

[0018] A4. The Canny edge detection algorithm is used to process the images in the net image set to extract the contour information of the transmission line components and obtain the edge image set;

[0019] A5. Calculate the image sharpness score based on the image contour information in the edge image set and compare it with a second preset threshold; for images whose sharpness reaches the second preset threshold, combine them with the corresponding environmental sensor data to construct the first round initial dataset; for images whose sharpness is lower than the second preset threshold, perform super-resolution reconstruction through a deep convolutional network, and after the sharpness reaches the second preset threshold, combine them with the corresponding environmental sensor data and include them in the first round initial dataset.

[0020] Step S3, which involves identifying anomalous images in the first round of the initial dataset to form an anomalous image dataset, includes the following steps:

[0021] B1. Extract environmental sensor data from the initial dataset of the first round to form a set of environmental condition parameters including light intensity, humidity and temperature;

[0022] B2. Based on the environmental condition parameter set, the sharpness score of the images in the first round initial dataset is calculated through a dynamic evaluation model; the sharpness score of the images in the first round initial dataset is compared with the third preset threshold, and images with a sharpness lower than the third preset threshold are marked as initial screening abnormal images;

[0023] B3. Perform image enhancement processing on the initial screening abnormal images and recalculate the sharpness score through a dynamic evaluation model; compare the sharpness score of the enhanced initial screening abnormal images with a third preset threshold, and select the initial screening abnormal images whose sharpness is still lower than the third preset threshold and mark them as abnormal images.

[0024] B4. Combine the abnormal images with the environmental sensor data corresponding to the abnormal images to form an abnormal image dataset.

[0025] Step S4, which involves determining the type of environmental factor causing the abnormal image based on the data content of the abnormal image dataset, includes the following steps:

[0026] C1. Compare the sharpness of each abnormal image in the abnormal image dataset with a fourth preset threshold; for abnormal images whose sharpness reaches the fourth preset threshold, extract their environmental feature data through a feature extraction algorithm to form an environmental feature dataset; the environmental feature dataset includes light intensity, color temperature, and texture details;

[0027] C2. For abnormal images with a resolution lower than the fourth preset threshold, image enhancement processing is performed first, and then their environmental feature data is extracted and included in the environmental feature dataset.

[0028] C3. Classify the environmental feature dataset using support vectors and algorithms to determine the types of environmental factors that cause abnormal images.

[0029] Step S5, which involves formulating corresponding sensor data acquisition tasks based on the type of environmental factors causing the abnormal images and assigning these tasks to various sensors to perform the second round of sensor data acquisition, includes the following steps:

[0030] D1. Based on the type of environmental factors that cause abnormal images, formulate corresponding sensor data acquisition tasks;

[0031] D2. According to the data transmission protocol, the sensor data acquisition task is transmitted to the collaborative scheduling system;

[0032] D3. A collaborative scheduling algorithm is used to analyze the resource status of each sensor, and based on the analysis results, sensor data acquisition tasks are assigned to each sensor.

[0033] The sensors mentioned in step S6 include vibration sensors and weather sensors. Acquiring the sensor data obtained in the second round of data collection and fusing it to obtain the initial dataset for the second round includes the following steps:

[0034] E1. Vibration sensors collect vibration data in real time to obtain vibration intensity and frequency information; meteorological sensors collect meteorological data in real time to obtain temperature, humidity and wind speed information.

[0035] E2. Synchronize the collection timestamps of vibration data and meteorological data, and merge the data to generate the second round of initial dataset.

[0036] Step S7, which involves optimizing and security processing the second round of initial dataset to generate a second round of secure data packets and complete the transmission, includes the following steps:

[0037] F1. The Kalman filter algorithm is used to adaptively filter the initial dataset of the second round to obtain the net dataset of the second round. Based on the accuracy requirements, the dataset is filtered to output the second round accurate dataset that meets the accuracy requirements.

[0038] F2. Obtain the second round of precision dataset through signal monitoring equipment, and extract the wireless communication signal strength features using time-domain analysis methods to obtain the signal strength values ​​of the second round of precision dataset;

[0039] F3. Compare the signal strength value with the fifth preset threshold, and select the second round of precision dataset whose signal strength value exceeds the fifth preset threshold as the second round of data to be encrypted;

[0040] F4. Encrypt the data using a preset encryption standard to obtain the second-round encrypted dataset;

[0041] F5. Extract data packets from the second-round encrypted dataset and generate second-round block-encrypted data packets using block encryption.

[0042] F6. Perform integrity verification on the second round of packet encryption data packets, select complete and protected data packets as the second round of secure data packets, and output them through the wireless communication module.

[0043] Step S8, which involves decrypting the received second-round security data packet to obtain the corresponding inspection data for the transmission line equipment, includes the following steps:

[0044] G1. Use a decryption algorithm to decrypt the second round of secure data packets and extract the second round of decrypted data packets;

[0045] G2. For the second round of decrypted data packets, data cleaning techniques are used to remove noise, resulting in the second round of clean data packets;

[0046] G3. The support vector machine algorithm is used to extract feature vectors from the second round of net data packets to obtain the state feature set of the transmission line, that is, the inspection data of the transmission line.

[0047] Secondly, a transmission line inspection device is provided, characterized in that the transmission line inspection device comprises: a first-round data acquisition module, a first-round data acquisition optimization module, an abnormal image recognition module, an environmental factor type analysis module, a sensor data acquisition task formulation, assignment, and execution module, a second-round sensor data fusion module, a second-round initial dataset optimization security module, and a second-round security data packet decryption module; the first-round data acquisition module, the first-round data acquisition optimization module, the abnormal image recognition module, the environmental factor type analysis module, the sensor data acquisition task formulation, assignment, and execution module, the second-round sensor data fusion module, the second-round initial dataset optimization security module, and the second-round security data packet decryption module are connected in series.

[0048] The first-round data acquisition module is used to acquire the first-round data acquisition dataset of the transmission line obtained in the first round of acquisition, including image data of the transmission line and environmental sensor data corresponding to the image, and upload the data to the optimization processing module of the first-round data acquisition dataset;

[0049] The optimization processing module for the first round of data collection is used to optimize the first round of data collection based on the received data, generate the first round initial dataset, and upload the data to the abnormal image recognition module.

[0050] The abnormal image recognition module is used to construct an abnormal image dataset based on the received data and the abnormal images in the initial dataset, and then upload the data to the environmental factor type analysis module.

[0051] The environmental factor type analysis module is used to determine the type of environmental factor causing the abnormal image based on the received data and the data content of the abnormal image dataset, and then upload the data to the sensor data acquisition task formulation and assignment module.

[0052] The sensor data acquisition task formulation, assignment, and execution module is used to formulate corresponding sensor data acquisition tasks based on the received data and the type of environmental factors that cause abnormal images, and assign the tasks to each sensor to execute the second round of sensor data acquisition, and upload the data to the second round of sensor data fusion module;

[0053] The second-round sensor data fusion module is used to fuse the received data and the sensor data acquired in the second round of acquisition to obtain the second-round initial dataset, and then upload the data to the second-round initial dataset optimization security module.

[0054] The second-round initial dataset optimization security module is used to optimize and securely process the second-round initial dataset based on the received data, generate a second-round secure data packet and complete its transmission, and then upload the data to the second-round secure data packet decryption module.

[0055] The second-round security data packet decryption module is used to decrypt the received second-round security data packet to obtain the corresponding inspection data of the transmission line equipment.

[0056] This application, through multiple rounds of data acquisition, can promptly identify anomalies in the currently collected data and, based on these anomalies, quickly discover environmental factors during transmission line inspection. This allows for targeted planning of the next round of sensor data acquisition tasks and timely correction of the collected data. Therefore, the transmission line inspection method and equipment provided in this application can better adapt to complex terrain and variable weather conditions, more accurately reflect the actual operating status of transmission lines, improve inspection efficiency, and promptly detect equipment faults. Attached Figure Description

[0057] Figure 1 This is a schematic diagram of the method flow of the present invention.

[0058] Figure 2 This is a schematic diagram of the functional modules of the device of the present invention. Detailed Implementation

[0059] like Figure 1 The diagram shown is a flowchart of the method of the present invention: The method for inspecting power transmission lines disclosed in this invention includes the following steps:

[0060] S1. Obtain the first round of data collection dataset of the transmission line obtained in the first round of data collection, including images of the transmission line and environmental sensor data corresponding to the images;

[0061] Images of power transmission lines include both infrared and visible light images, acquired simultaneously by an infrared thermal imager and a visible light camera. The infrared thermal imager captures abnormally high temperatures generated by overload at the transmission line joints, while the visible light camera records the physical morphology of the joints, such as corrosion or deformation. Simultaneous acquisition ensures that the two images are aligned in time and space, providing a reliable basis for subsequent image fusion. For example, the infrared image may show a joint temperature of 80°C, exceeding the normal range by 50°C, while the visible light image may reveal surface cracks. This combination of dual-modal data provides a comprehensive basis for fault diagnosis.

[0062] S2. Optimize the first round of collected datasets to generate the first round of initial datasets; specifically including the following steps:

[0063] A1. A weighted average fusion algorithm is used to perform pixel-level fusion of infrared and visible light images in the transmission line images to form a fused image set.

[0064] This fusion aims to integrate the thermal information of the infrared image with the detailed information of the visible light image. Weighted average fusion balances the contributions of the two images by assigning different weights.

[0065] For example, the infrared image is weighted at 0.6 to highlight temperature anomalies, while the visible light image is weighted at 0.4 to preserve structural details. The fused image clearly shows the correspondence between the high-temperature region of the joint and the crack location, improving the visualization of the defect. This method is computationally simple and retains key information, which is helpful for subsequent analysis.

[0066] A2. Using histogram equalization, the images in the fused image set are enhanced by adjusting the image contrast to obtain the enhanced image set;

[0067] For example, fused images may suffer from low contrast due to insufficient light or weak thermal signals. Histogram equalization makes dark, high-temperature areas more prominent by redistributing pixel grayscale values. After processing, the grayscale difference between the joint crack and the high-temperature area increased from 20 to 50, significantly improving visual recognition. This method effectively improves image quality and provides better input for subsequent edge detection.

[0068] A3. Compare the pixel value variance of the images in the enhanced image set with a first preset threshold, filter out images whose pixel value variance is lower than the first preset threshold, and use median filtering to denoise them to obtain a net image set;

[0069] For example, the first preset threshold is set to 100. Random noise in transmission line images may interfere with contour extraction. Median filtering calculates the median in a 3x3 neighborhood, smoothing noise points while preserving joint edges. This method can improve image clarity and reduce the risk of false detections in low-variance scenarios.

[0070] A4. The Canny edge detection algorithm is used to process the images in the net image set to extract the contour information of the transmission line components and obtain the edge image set;

[0071] The Canny algorithm accurately identifies edges through gradient calculation and double thresholding. For example, the boundaries of joint cracks and high-temperature areas are clearly delineated, and the edge intensity is enhanced from a blurred transition in the fused image to a clear outline.

[0072] Preferably, the dual thresholds are set to 50 and 150 to ensure that weak edges are preserved and strong edges are not lost. This method is highly effective in extracting the contours of power transmission components against complex backgrounds, providing reliable data for sharpness assessment.

[0073] A5. Calculate the image sharpness score based on the image contour information in the edge image set and compare it with a second preset threshold; for images whose sharpness reaches the second preset threshold, combine them with the corresponding environmental sensor data to construct the first round initial dataset; for images whose sharpness is lower than the second preset threshold, perform super-resolution reconstruction through a deep convolutional network, and after the sharpness reaches the second preset threshold, combine them with the corresponding environmental sensor data and include them in the first round initial dataset.

[0074] For example, the sharpness score is based on edge sharpness and contour continuity, with a scoring range of 0-100, and the second preset threshold is set to 80. If the sharpness score of the power transmission line connector image is 85, reaching the second preset threshold, then the power transmission line connector image and the corresponding environmental sensor data are combined and included in the first round of initial dataset. If the sharpness score of the power transmission line connector image is lower than 80, then the power transmission line connector image is super-resolution reconstructed using a deep convolutional network. For example, if the resolution is increased from 256x256 to 512x512, the sharpness reaches the second preset threshold, and the power transmission line connector image is combined with the corresponding environmental sensor data and included in the first round of initial dataset.

[0075] By optimizing the characteristics of transmission line images, the images can effectively support fault detection and maintenance decisions, significantly improving diagnostic efficiency and accuracy.

[0076] S3. Identify anomalous images in the initial dataset of the first round to form an anomalous image dataset; this includes the following steps:

[0077] B1. Extract environmental sensor data from the initial dataset of the first round to form a set of environmental condition parameters including light intensity, humidity and temperature;

[0078] Illumination intensity was estimated using the average grayscale value of the abnormal images, humidity was obtained by combining on-site sensor data, and temperature was extracted from the thermal information of the infrared images.

[0079] For example, a dataset of images of power line connectors shows an illumination intensity of 200 lux, humidity of 75%, and temperature of 45°C. These parameters reflect the environmental conditions at the time of image acquisition and provide a basis for subsequent sharpness assessment.

[0080] B2. Based on the environmental condition parameter set, the sharpness score of the images in the first round initial dataset is calculated through a dynamic evaluation model; the sharpness score of the images in the first round initial dataset is compared with the third preset threshold, and images with a sharpness lower than the third preset threshold are marked as initial screening abnormal images;

[0081] For example, the dynamic evaluation model generates a sharpness score by weighting the effects of illumination and humidity, combined with the effect of temperature on image contrast. If a connector image has a score of 70 due to insufficient illumination, which is below the third preset threshold of 80, it is judged as an abnormal image in the initial screening.

[0082] B3. Perform image enhancement processing on the initial screening abnormal images and recalculate the sharpness score through a dynamic evaluation model; compare the sharpness score of the enhanced initial screening abnormal images with a third preset threshold, and select the initial screening abnormal images whose sharpness is still lower than the third preset threshold and mark them as abnormal images.

[0083] Image enhancement processing can be achieved through contrast stretching or sharpening filtering. Contrast stretching enhances joint details by expanding the pixel grayscale range. For example, expanding the grayscale range of a power transmission line image from 50-150 to 0-255 makes the surface texture of the joint more apparent. Sharpening filtering, on the other hand, highlights contours by enhancing high-frequency information at the edges. The resulting enhanced image has clearer details. If the sharpness score of the enhanced joint image is still below a third preset threshold of 80, it is marked as an abnormal image.

[0084] Preferably, a list containing abnormal image identifiers is generated based on the abnormal image.

[0085] For example, anomaly image generation can create a list containing anomaly image identifiers. Suppose a transmission line dataset contains 1000 images; after filtering, 10 images score below 80. The list records their numbers, such as "IMG_0231, IMG_0456". This list provides precise location information for subsequent maintenance. Once generated, the list can be directly used to guide on-site inspections, quickly identifying joint images requiring further processing.

[0086] B4. Combine the abnormal images with the environmental sensor data corresponding to the abnormal images to form an abnormal image dataset.

[0087] For example, "Abnormal image IMG_0231, illumination 150 lux, humidity 80%, temperature 50°C" makes it easier to analyze the reasons for insufficient clarity.

[0088] By using multi-layered filtering based on image clarity, abnormal images in the first round of data collection can be identified relatively accurately, providing support for subsequent fault finding in power transmission lines and correction of collected sensor data.

[0089] S4. Based on the data content of the abnormal image dataset, determine the type of environmental factors that caused the abnormal images; this includes the following steps:

[0090] C1. Compare the sharpness of each abnormal image in the abnormal image dataset with a fourth preset threshold; for abnormal images whose sharpness reaches the fourth preset threshold, extract their environmental feature data through a feature extraction algorithm to form an environmental feature dataset; the environmental feature dataset includes light intensity, color temperature, and texture details;

[0091] C2. For abnormal images with a resolution lower than the fourth preset threshold, image enhancement processing is performed first, and then their environmental feature data is extracted and included in the environmental feature dataset.

[0092] Suppose we have a dataset containing 1000 images, of which 200 images have a sharpness value below a fourth preset threshold of 0.7. We employ adaptive histogram equalization to enhance the image contrast.

[0093] For example, for images with insufficient lighting, adjusting the brightness distribution makes details in dark areas more apparent. The core of this method lies in improving the image's visualization by adjusting the pixel value distribution, thus providing clearer input for subsequent feature extraction.

[0094] C3. Classify the environmental feature dataset using support vectors and algorithms to determine the types of environmental factors that cause abnormal images.

[0095] Assuming the environmental feature dataset includes illumination, color temperature, and texture features, the environmental factor types are divided into three categories: "insufficient illumination," "abnormal color temperature," and "blurred texture."

[0096] For example, by training a support vector machine model and setting the kernel function to radial basis function, the classification results show that 70% of the images belong to the under-illuminated type.

[0097] S5. Based on the type of environmental factors causing the abnormal images, formulate corresponding sensor data acquisition tasks and assign these tasks to each sensor to perform the second round of sensor data acquisition; specifically, this includes the following steps:

[0098] D1. Based on the type of environmental factors that cause abnormal images, formulate corresponding sensor data acquisition tasks;

[0099] Assuming that the majority of images are underlit, the sensor data acquisition tasks can include "adding supplementary lighting" and "adjusting white balance settings." These tasks are encapsulated in JSON format, containing the task ID, priority, and execution time.

[0100] For example, the supplemental lighting task with task ID T001 has a high priority and a planned execution time of 10 minutes. This structured task list facilitates system parsing and execution.

[0101] D2. According to the data transmission protocol, the sensor data acquisition task is transmitted to the collaborative scheduling system;

[0102] Low-latency transmission can be achieved by using the MQTT protocol.

[0103] D3. A collaborative scheduling algorithm is used to analyze the resource status of each sensor, and based on the analysis results, sensor data acquisition tasks are assigned to each sensor.

[0104] For example, consider an environmental monitoring scenario involving power transmission lines at a wind farm. The system deploys 10 vibration sensors and 5 weather sensors. Based on a collaborative scheduling algorithm, sensor data acquisition tasks are assigned to each sensor according to its remaining power, signal strength, and sampling frequency.

[0105] Preferably, the algorithm will prioritize assigning tasks to sensors with sufficient power and stable signals.

[0106] For example, vibration sensor 1, being closer to the control center, has a signal strength of 95% and a battery level of 80%, so it will be assigned a higher sampling frequency, such as 10 times per second, while sensor 5, with a weaker signal, will be assigned a lower frequency, such as 2 times per second, in order to balance resource usage.

[0107] S6. Acquire the sensor data obtained in the second round of collection, fuse and process them to obtain the initial dataset for the second round; the sensors include vibration sensors and meteorological sensors; specifically, the following steps are included:

[0108] E1. Vibration sensors collect vibration data in real time to obtain vibration intensity and frequency information; meteorological sensors collect meteorological data in real time to obtain temperature, humidity and wind speed information.

[0109] For example, in wind farms, vibration sensors are installed on the blades and towers of wind turbines.

[0110] The sensor detected a blade vibration intensity of 0.2g and a frequency of 5Hz, indicating a potential blade imbalance. Meteorological sensors recorded a temperature of 20°C, humidity of 60%, and wind speed of 8m / s; these data reflect the potential impact of the current environment on the wind turbine's operation.

[0111] E2. Synchronize the collection timestamps of vibration data and meteorological data, and merge the data to generate the second round of initial dataset.

[0112] For example, by using network time protocols, the timestamps of vibration and meteorological data are accurate to the millisecond level, ensuring the accuracy of data fusion. Correlation analysis was performed between vibration intensity of 0.2g, frequency of 5Hz, wind speed of 8m / s, and humidity of 60% to generate the second round of initial datasets.

[0113] S7. Optimize and securely process the initial dataset for the second round, generate a secure data packet for the second round, and complete the transmission; specifically, this includes the following steps:

[0114] F1. The Kalman filter algorithm is used to adaptively filter the initial dataset of the second round to obtain the net dataset of the second round. Based on the accuracy requirements, the dataset is filtered to output the second round accurate dataset that meets the accuracy requirements.

[0115] Vibration sensors may be deployed near industrial equipment to capture vibration signals from mechanical operation. The initial signals often contain noise, such as the superposition of the equipment's own vibration and external environmental interference. Weather sensors record parameters such as temperature, humidity, and wind speed. The data may be introduced with noise due to sudden environmental changes, so the initial dataset needs to be denoised in the second round.

[0116] For example, after filtering, the noise amplitude of the vibration signal decreases from 0.5 mm / s to 0.1 mm / s, resulting in a smoother signal. In meteorological data, humidity spikes are filtered out, and the data stream tends to stabilize. Filtering parameters, such as the initial process noise covariance of 0.01 and the measurement noise covariance of 0.1, need to be adjusted according to the actual noise characteristics.

[0117] When calculating the noise suppression rate, the noise amplitude of the signal before and after filtering is compared. A noise suppression rate of 80% for vibration signals indicates that most interference is eliminated. In meteorological data, the suppression rate is approximately 75% after filtering out humidity spikes. To ensure the data stream meets the accuracy requirements of the application scenario, the data stream's suitability for downstream tasks, such as equipment failure prediction, is analyzed in conjunction with sensor accuracy (e.g., vibration sensor accuracy ±0.02 mm / s, meteorological sensor accuracy ±0.5°C). If vibration data fluctuations are controlled within ±0.03 mm / s and temperature data errors are less than ±0.4°C, then the industrial monitoring requirements are met.

[0118] Preferably, before outputting the second round of precision dataset in this embodiment, it is necessary to verify whether the data acquisition frequency supports real-time applications.

[0119] For example, a vibration sensor with a frequency of 100 Hz is suitable for capturing instantaneous vibrations, while a weather sensor with a frequency of 1 Hz is sufficient to reflect environmental changes.

[0120] F2. Obtain the second round of precision dataset through signal monitoring equipment, and extract the wireless communication signal strength features using time-domain analysis methods to obtain the signal strength values ​​of the second round of precision dataset;

[0121] A clean data stream is obtained by acquiring environmental signals at a sampling rate of 100Hz using a high-precision sensor. Time-domain analysis methods are used to extract signal strength features, which can be characterized by calculating the peak amplitude or average power of the signal.

[0122] F3. Compare the signal strength value with the fifth preset threshold, and select the second round of precision dataset whose signal strength value exceeds the fifth preset threshold as the second round of data to be encrypted;

[0123] If the signal strength is -50dBm and the fifth preset threshold is -60dBm, and the signal strength has exceeded the fifth preset threshold, then there is a risk of leakage, and it should be included in the second round of data to be encrypted.

[0124] Preferably, a risk identifier is generated after determining that there is a risk of data leakage. The risk identifier is a simple binary flag, such as 1 indicating high risk and 0 indicating safe.

[0125] For example, if the monitoring equipment analyzes the signal strength and five consecutive samples exceed -60 dBm, a high-risk flag is generated. This flag generation is based on statistical analysis to ensure a low false alarm rate.

[0126] F4. Encrypt the data using a preset encryption standard to obtain the second-round encrypted dataset;

[0127] When activating the encryption protocol, the default encryption standard is typically AES-256, with a key length of 256 bits. For example, the data stream is encrypted using AES-256 to generate a second-round encrypted data set. Every 128-bit data block is XORed with the key and then replaced to ensure data security.

[0128] F5. Extract data packets from the second-round encrypted dataset and generate second-round block-encrypted data packets using block encryption.

[0129] When extracting data packets from the second-round encrypted dataset, the block cipher method uses the CBC mode. The second-round encrypted dataset is divided into fixed-size data packets, such as 1024 bytes, and encrypted after adding an initialization vector. The second-round encrypted dataset is then divided into multiple data packets, each with a timestamp added as an identifier, and encrypted to generate the second-round block cipher data packets.

[0130] F6. Perform integrity verification on the second round of packet encryption data packets, select complete and protected data packets as the second round of secure data packets, and output them through the wireless communication module.

[0131] For example, the hash value of the second round of packet encryption data is calculated and compared with the receiving end. If they match, the second round of packet encryption data is considered complete and can be output through a wireless communication module such as Wi-Fi or 5G protocol.

[0132] Based on the transmission logs, the transmission status of data packets is analyzed to determine if any anomalies exist, resulting in a transmission status analysis. For example, if a device uses a 5G module to transmit data packets at a rate of 1Gbps, the log records show a transmission latency of less than 10ms. The transmission status analysis is based on checking the packet loss rate or the number of retransmissions in the logs.

[0133] For example, if the logs show a packet loss rate of less than 0.1%, the transmission status is considered normal.

[0134] S8. Decrypt the received second-round security data packet to obtain the corresponding inspection data of the transmission line equipment, which includes the following steps;

[0135] G1. Use a decryption algorithm to decrypt the second round of secure data packets and extract the second round of decrypted data packets;

[0136] For example, environmental data collected by sensor nodes is encrypted using AES-128 to form data packets. Decryption uses a pre-shared key to extract the original state dataset containing data such as temperature, humidity, and vibration frequency. The decryption process ensures that the data has not been tampered with, and its integrity is verified through checksums.

[0137] For example, after decrypting a data packet from an industrial sensor, a dataset with a temperature of 28.5°C, humidity of 60%, and vibration frequency of 50Hz is obtained, which conforms to the preset JSON format, indicating that the data is complete.

[0138] G2. For the second round of decrypted data packets, data cleaning techniques are used to remove noise, resulting in the second round of clean data packets;

[0139] Specifically, data cleaning techniques are used to remove noisy data. Sensor data may contain outliers due to electromagnetic interference, such as sudden temperature jumps to 100°C. Median filtering is used to remove data outside the normal range, retaining reasonable temperature values ​​between 28.0°C and 29.0°C, resulting in a cleaned state dataset. Cleaning ensures data accuracy and provides a reliable foundation for subsequent analysis.

[0140] G3. The support vector machine algorithm is used to extract feature vectors from the second round of net data packets to obtain the state feature set of the transmission line, that is, the inspection data of the transmission line.

[0141] For example, features such as temperature change rate, humidity fluctuation amplitude, and vibration frequency stability can be extracted from the cleaned dataset to form a state feature set of the transmission line, i.e., the inspection data of the transmission line.

[0142] The transmission line inspection method and equipment provided in this application can better adapt to complex terrain and variable climate conditions, more accurately reflect the actual operating status of transmission lines, improve inspection efficiency, and detect equipment failures in a timely manner.

[0143] like Figure 2The diagram shows the functional modules of the device of the present invention: The transmission line inspection device disclosed in this invention includes: a first-round data acquisition module, a first-round data acquisition optimization processing module, an abnormal image recognition module, an environmental factor type analysis module, a sensor data acquisition task formulation, assignment and execution module, a second-round sensor data fusion module, a second-round initial dataset optimization security module, and a second-round security data packet decryption module; the first-round data acquisition module, the first-round data acquisition optimization processing module, the abnormal image recognition module, the environmental factor type analysis module, the sensor data acquisition task formulation, assignment and execution module, the second-round sensor data fusion module, the second-round initial dataset optimization security module, and the second-round security data packet decryption module are connected in series;

[0144] The first-round data acquisition module is used to acquire the first-round data acquisition dataset of the transmission line obtained in the first round of acquisition, including image data of the transmission line and environmental sensor data corresponding to the image, and upload the data to the optimization processing module of the first-round data acquisition dataset;

[0145] The optimization processing module for the first round of data collection is used to optimize the first round of data collection based on the received data, generate the first round initial dataset, and upload the data to the abnormal image recognition module.

[0146] The abnormal image recognition module is used to construct an abnormal image dataset based on the received data and the abnormal images in the initial dataset, and then upload the data to the environmental factor type analysis module.

[0147] The environmental factor type analysis module is used to determine the type of environmental factor causing the abnormal image based on the received data and the data content of the abnormal image dataset, and then upload the data to the sensor data acquisition task formulation and assignment module.

[0148] The sensor data acquisition task formulation, assignment, and execution module is used to formulate corresponding sensor data acquisition tasks based on the received data and the type of environmental factors that cause abnormal images, and assign the tasks to each sensor to execute the second round of sensor data acquisition, and upload the data to the second round of sensor data fusion module;

[0149] The second-round sensor data fusion module is used to fuse the received data and the sensor data acquired in the second round of acquisition to obtain the second-round initial dataset, and then upload the data to the second-round initial dataset optimization security module.

[0150] The second-round initial dataset optimization security module is used to optimize and securely process the second-round initial dataset based on the received data, generate a second-round secure data packet and complete its transmission, and then upload the data to the second-round secure data packet decryption module.

[0151] The second-round security data packet decryption module is used to decrypt the received second-round security data packet to obtain the corresponding inspection data of the transmission line equipment.

Claims

1. A method for inspecting power transmission lines, characterized in that, Includes the following steps: S1. Obtain the first round of data collection dataset of the transmission line obtained in the first round of data collection, including images of the transmission line and environmental sensor data corresponding to the images; S2. Optimize the first round of collected datasets to generate the first round of initial datasets; S3. Identify anomalous images in the initial dataset of the first round to form an anomalous image dataset; S4. Based on the data content of the abnormal image dataset, determine the type of environmental factors that caused the abnormal images; S5. Based on the type of environmental factors that cause abnormal images, formulate corresponding sensor data acquisition tasks and assign these tasks to each sensor to perform the second round of sensor data acquisition. S6. Acquire the sensor data obtained in the second round of acquisition, fuse and process it to obtain the initial dataset for the second round; S7. Optimize and securely process the initial dataset for the second round, generate a secure data packet for the second round, and complete the transmission. S8. Decrypt the received second-round security data packet to obtain the corresponding inspection data of the transmission line equipment.

2. The method for inspecting transmission lines according to claim 1, characterized in that, Step S2 includes the following steps: A1. A weighted average fusion algorithm is used to perform pixel-level fusion of infrared and visible light images in the images of transmission lines to form a fused image set; A2. Using histogram equalization, the images in the fused image set are enhanced by adjusting the image contrast to obtain the enhanced image set; A3. Compare the pixel value variance of the images in the enhanced image set with a first preset threshold, filter out images whose pixel value variance is lower than the first preset threshold, and use median filtering to denoise them to obtain a net image set; A4. The Canny edge detection algorithm is used to process the images in the net image set to extract the contour information of the transmission line components and obtain the edge image set; A5. Calculate the image sharpness score based on the image contour information in the edge image set and compare it with a second preset threshold; for images whose sharpness reaches the second preset threshold, combine them with the corresponding environmental sensor data to construct the first round initial dataset; for images whose sharpness is lower than the second preset threshold, perform super-resolution reconstruction through a deep convolutional network, and after the sharpness reaches the second preset threshold, combine them with the corresponding environmental sensor data and include them in the first round initial dataset.

3. The method for inspecting transmission lines according to claim 2, characterized in that, Step S3 includes the following steps: B1. Extract environmental sensor data from the initial dataset of the first round to form a set of environmental condition parameters including light intensity, humidity and temperature; B2. Based on the environmental condition parameter set, the sharpness score of the images in the first round initial dataset is calculated through a dynamic evaluation model; the sharpness score of the images in the first round initial dataset is compared with the third preset threshold, and images with a sharpness lower than the third preset threshold are marked as initial screening abnormal images; B3. Perform image enhancement processing on the initial screening abnormal images and recalculate the sharpness score through a dynamic evaluation model; compare the sharpness score of the enhanced initial screening abnormal images with a third preset threshold, and select the initial screening abnormal images whose sharpness is still lower than the third preset threshold and mark them as abnormal images. B4. Combine the abnormal images with the environmental sensor data corresponding to the abnormal images to form an abnormal image dataset.

4. The method for inspecting transmission lines according to claim 3, characterized in that, Step S4 includes the following steps: C1. Compare the sharpness of each abnormal image in the abnormal image dataset with a fourth preset threshold; for abnormal images whose sharpness reaches the fourth preset threshold, extract their environmental feature data through a feature extraction algorithm to form an environmental feature dataset; the environmental feature dataset includes light intensity, color temperature, and texture details; C2. For abnormal images with a resolution lower than the fourth preset threshold, image enhancement processing is performed first, and then their environmental feature data is extracted and included in the environmental feature dataset. C3. Classify the environmental feature dataset using support vectors and algorithms to determine the types of environmental factors that cause abnormal images.

5. The method for inspecting transmission lines according to claim 4, characterized in that, Step S5 includes the following steps: D1. Based on the type of environmental factors that cause abnormal images, formulate corresponding sensor data acquisition tasks; D2. According to the data transmission protocol, the sensor data acquisition task is transmitted to the collaborative scheduling system; D3. A collaborative scheduling algorithm is used to analyze the resource status of each sensor, and based on the analysis results, sensor data acquisition tasks are assigned to each sensor.

6. The method for inspecting transmission lines according to claim 5, characterized in that, The sensors mentioned in step S6 include vibration sensors and weather sensors. Acquiring the sensor data obtained in the second round of data collection and fusing it to obtain the initial dataset for the second round includes the following steps: E1. Vibration sensors collect vibration data in real time to obtain vibration intensity and frequency information; meteorological sensors collect meteorological data in real time to obtain temperature, humidity and wind speed information. E2. Synchronize the collection timestamps of vibration data and meteorological data, and merge the data to generate the second round of initial dataset.

7. The method for inspecting transmission lines according to claim 6, characterized in that, Step S7 includes the following steps: F1. The Kalman filter algorithm is used to adaptively filter the initial dataset of the second round to obtain the net dataset of the second round. Based on the accuracy requirements, the dataset is filtered to output the second round accurate dataset that meets the accuracy requirements. F2. Obtain the second round of precision dataset through signal monitoring equipment, and extract the wireless communication signal strength features using time-domain analysis methods to obtain the signal strength values ​​of the second round of precision dataset; F3. Compare the signal strength value with the fifth preset threshold, and select the second round of precision dataset whose signal strength value exceeds the fifth preset threshold as the second round of data to be encrypted; F4. Encrypt the data using a preset encryption standard to obtain the second-round encrypted dataset; F5. Extract data packets from the second-round encrypted dataset and generate second-round block-encrypted data packets using block encryption. F6. Perform integrity verification on the second round of packet encryption data packets, select complete and protected data packets as the second round of secure data packets, and output them through the wireless communication module.

8. The method for inspecting transmission lines according to claim 7, characterized in that, Step S8 includes the following steps: G1. Use a decryption algorithm to decrypt the second round of secure data packets and extract the second round of decrypted data packets; G2. For the second round of decrypted data packets, data cleaning techniques are used to remove noise, resulting in the second round of clean data packets; G3. The support vector machine algorithm is used to extract feature vectors from the second round of net data packets to obtain the state feature set of the transmission line, that is, the inspection data of the transmission line.

9. An inspection device for power transmission lines, characterized in that, The transmission line inspection equipment includes: a first-round dataset acquisition module, a first-round dataset optimization and processing module, an abnormal image recognition module, an environmental factor type analysis module, a sensor data acquisition task formulation, assignment, and execution module, a second-round sensor data fusion module, a second-round initial dataset optimization and security module, and a second-round security data packet decryption module; the first-round dataset acquisition module, the first-round dataset optimization and processing module, the abnormal image recognition module, the environmental factor type analysis module, the sensor data acquisition task formulation, assignment, and execution module, the second-round sensor data fusion module, the second-round initial dataset optimization and security module, and the second-round security data packet decryption module are connected in series. The first-round data acquisition module is used to acquire the first-round data acquisition dataset of the transmission line obtained in the first round of acquisition, including image data of the transmission line and environmental sensor data corresponding to the image, and upload the data to the optimization processing module of the first-round data acquisition dataset; The optimization processing module for the first round of data collection is used to optimize the first round of data collection based on the received data, generate the first round initial dataset, and upload the data to the abnormal image recognition module. The abnormal image recognition module is used to construct an abnormal image dataset based on the received data and the abnormal images in the initial dataset, and then upload the data to the environmental factor type analysis module. The environmental factor type analysis module is used to determine the type of environmental factor causing the abnormal image based on the received data and the data content of the abnormal image dataset, and then upload the data to the sensor data acquisition task formulation and assignment module. The sensor data acquisition task formulation, assignment, and execution module is used to formulate corresponding sensor data acquisition tasks based on the received data and the type of environmental factors that cause abnormal images, and assign the tasks to each sensor to execute the second round of sensor data acquisition, and upload the data to the second round of sensor data fusion module; The second-round sensor data fusion module is used to fuse the received data and the sensor data acquired in the second round of acquisition to obtain the second-round initial dataset, and then upload the data to the second-round initial dataset optimization security module. The second-round initial dataset optimization security module is used to optimize and securely process the second-round initial dataset based on the received data, generate a second-round secure data packet and complete its transmission, and then upload the data to the second-round secure data packet decryption module. The second-round security data packet decryption module is used to decrypt the received second-round security data packet to obtain the corresponding inspection data of the transmission line equipment.

Images