Power line data acquisition device and method

Abstract

The invention relates to the technical field of electric power circuit data acquisition, and provides an electric power circuit data acquisition device and method, which can acquire related data of an electric power circuit and peripheral equipment thereof in real time through the combination of a laser range finder and a photographing module, including the distances from the ground, houses and trees and the phase sequence information of the circuit. The deep learning model is used for analysis, and the distance between the hidden danger safety line and the adjacent equipment can be accurately judged, so that the potential hidden danger is recognized in advance, the emergency repair requirement after the fault occurs is reduced, and the customer complaint risk is reduced. The problem of line phase sequence errors is effectively solved, and the phase sequence errors can be found and corrected in time through real-time monitoring and analysis of the phase sequence information, so that the phenomena of heavy load and low voltage of a distribution transformer district are avoided, and the stability of power supply is ensured. The data preprocessing stage comprises data cleaning, denoising and standardization. New data are strictly verified through an incremental data verification layer, and it is guaranteed that only compliant data are input into a system.

Description

Technical Field

[0001] The present invention relates to the technical field of power line data acquisition, and in particular to a power line data acquisition device and method. Background Art

[0002] In the era of rapid development of informatization, data has become an important resource for various industries and is regarded as a national strategic resource. As a core component of energy enterprises, power supply bureaus are increasingly dependent on data, and the effective use of data is crucial to ensuring power supply safety. Through in-depth analysis of power line data, power supply bureaus can promptly discover safety hazards on site, thereby taking corresponding rectification measures to avoid safety accidents and customer complaints.

[0003] Take Zhaotong Power Supply Bureau as an example. The bureau is responsible for managing 28,163 distribution transformers and 986 lines, and faces huge challenges in data collection and management. Due to the large number of distribution transformers and lines, there are many pain points in the collection and analysis of power line data:

[0004] 1. Insufficient identification of hidden dangers: Currently, the distance between hidden danger safety lines and adjacent equipment cannot be accurately judged, making it difficult to identify hidden dangers. Emergency repairs are only carried out after a fault occurs, which not only increases the risk of customer complaints, but may also cause personal injury accidents;

[0005] 2. Power supply problems caused by incorrect phase sequence: The incorrect phase sequence of the line will cause overload and low voltage in the distribution transformer area, which will affect the reasonable adjustment of the load and the stability of the power supply;

[0006] 3. Data entry relies on manual labor: Data entry work is closely related to the skill level of personnel, and data problems caused by subjective factors require a lot of manpower to correct.

[0007] Based on this, the present invention proposes a power line data acquisition device and method to improve the accuracy and efficiency of data acquisition, thereby improving power supply safety and service quality. Summary of the invention

[0008] The purpose of the present invention is to solve the deficiencies of the prior art and to provide a power line data acquisition device and method to improve the accuracy and efficiency of data acquisition, thereby improving power supply safety and service quality.

[0009] In order to achieve the above object, the present invention adopts the following technical solutions:

[0010] A power line data acquisition device, comprising: an acquisition module, a processing module, an analysis module, a human-computer interaction module and a verification module;

[0011] The acquisition module is used to collect power line related data in real time, including pole model, line length, cable width, distance from the ground, houses, trees and phase sequence information;

[0012] The processing module is used to preliminarily process and organize the information output by the acquisition module;

[0013] The analysis module uses a preset deep learning model to analyze the processed data, identify potential risks and provide decision support;

[0014] The human-computer interaction module is used to provide a friendly interface for operators, display analysis results in real time, and support user interaction;

[0015] The verification module is used to control the entry of incremental data and prevent the addition of abnormal data.

[0016] Preferably, the acquisition module includes: a laser rangefinder and a camera module; the laser rangefinder is used to collect the pole model, line length, distance to the ground when the line crosses or is close to ground roads, houses, trees, etc., proximity distance data, and cable width information; the camera module is used to collect images of power lines to be identified.

[0017] Preferably, the processing module performs data preprocessing and edge computing on the information output by the acquisition module; the data preprocessing includes data cleaning, denoising, and standardization; and the edge computing performs real-time processing and integration of the data.

[0018] Preferably, the analysis module includes: a data input layer, a deep learning model, a feature fusion layer, a hidden danger identification layer and a decision support layer; the data input layer is used to receive the integrated data output by the processing module and convert it into the input format required by the deep learning model;

[0019] The deep learning model includes a convolutional neural network (CNN) for image feature extraction and a long short-term memory (LSTM) network for time series data analysis;

[0020] The feature fusion layer fuses the features from the laser rangefinder and the camera module, forms a comprehensive feature vector through feature engineering technology, and improves the weight of key features through the attention mechanism;

[0021] The hidden danger identification layer classifies and regresses the fused features to identify potential safety hazards; the decision support layer generates decision suggestions based on the identification results.

[0022] Preferably, the verification module includes: a data receiving layer and an incremental data verification layer; the data receiving layer is responsible for receiving the integrated data from the processing module;

[0023] The incremental data verification layer verifies new data according to set rules and thresholds, and prevents the addition of abnormal data.

[0024] Another object of the present invention is to provide a method for collecting power line data, comprising the following steps:

[0025] S1. Data collection: The laser rangefinder and the camera module run simultaneously to collect information related to the power lines; the information related to the power lines includes: pole model, line length, cable width, distance from the ground, houses, trees, and phase sequence information;

[0026] S2, data preprocessing: In the processing module, the data output by the laser rangefinder is cleaned to remove outliers; the image obtained by the camera module is subjected to image denoising and enhancement processing, the brightness, contrast and saturation of the image are adjusted, and the image is cropped and scaled to a uniform size;

[0027] S3, edge computing: process the data output by the laser rangefinder and camera module, integrate them into a unified format, reduce data transmission delay, and compress data to improve transmission efficiency;

[0028] S4. Data analysis: Use the deep learning model preset in the analysis module to analyze the processed data, identify potential risks and send alarms in time;

[0029] S5. Decision support: Generate decision suggestions based on the analysis results, provide maintenance suggestions, maintenance plans and fault reports to operators in real time through the human-computer interaction module, and set early warning thresholds so that hidden dangers can be identified;

[0030] S6. Data verification: In the verification module, the incremental data is verified to ensure that only compliant data is entered into the system and to prevent the addition of abnormal data.

[0031] Preferably, the step S4 includes the following steps:

[0032] S4.1, data input, receiving the integrated data output by the processing module through the data input layer and converting it into the input format required by the deep learning model;

[0033] S4.2. Model selection and training. The deep learning framework TensorFlow is used for model building and training, including the use of cross-validation and hyperparameter tuning.

[0034] S4.3, feature extraction and fusion, use the trained deep learning model to extract features from the data, and analyze it by combining image features and ranging information; in the feature fusion layer, the features from the laser rangefinder and the camera module are fused, and feature engineering technology is used to form a comprehensive feature vector, and the weight of key features is increased through the attention mechanism;

[0035] S4.4. Hazard analysis: At the hazard identification layer, classification and regression analysis are performed on the fusion features to identify potential safety hazards.

[0036] Preferably, in step S4.3, a convolutional neural network is used to process image data to identify features in the image; a long short-term memory network is used to process time series data to analyze the changing trend of ranging information, and the weight of important features is increased in combination with an attention mechanism.

[0037] Preferably, in step S1, the laser rangefinder distance measurement algorithm requires the following inputs: two distances measured by the laser ranging module: LA: the distance from point C to point A; LB: the distance from point C to point B; the angle θ measured by the angle sensor, i.e. the angle between LA and LB at point C.

[0038] The present invention discloses a power line data acquisition device having the following beneficial effects.

[0039] The present invention uses a combination of a laser rangefinder and a camera module to obtain relevant data of power lines and their peripheral equipment in real time, including the distance from the ground, houses, trees, and the phase sequence information of the lines. By using a deep learning model for analysis, it is possible to accurately determine the distance between the hidden danger safety line and the adjacent equipment, thereby identifying potential hidden dangers in advance, reducing the need for emergency repairs after a fault occurs, reducing the risk of customer complaints, and ensuring personal safety. It effectively solves the problem of line phase sequence errors. Through real-time monitoring and analysis of phase sequence information, phase sequence errors can be discovered and corrected in a timely manner, thereby avoiding overload and low voltage phenomena in the distribution and transformation area, and ensuring the stability of power supply. The data preprocessing stage includes data cleaning, denoising, and standardization to ensure the high quality of the input data. Through the incremental data verification layer, new data is strictly verified to ensure that only compliant data is entered into the system, abnormal data is eliminated, and the data quality and stability of the system are effectively maintained. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 The figure is a system structure diagram of the power line data acquisition device of the present invention.

[0041] Figure 2 This is a data flow chart of the power line of the present invention.

[0042] Figure 3 It is a structural schematic diagram of the power line data acquisition device of the present invention.

[0043] Figure 4 It is a partial structural schematic diagram of the power line data acquisition device of the present invention.

[0044] 1-acquisition module; 2-processing module; 3-analysis module; 4-human-computer interaction module; 5-verification module; 6-lighting lamp; 7-housing; 8-operation button; 9-left and right locking levers; 10-left and right knobs; 11-front and rear locking levers; 12-front and rear knobs; 13-pan-tilt base plate; 14-display housing. DETAILED DESCRIPTION

[0045] In order to explain in detail the possible application scenarios, technical principles, specific schemes that can be implemented, and the purposes and effects that can be achieved, the following is a detailed description of the specific embodiments listed in conjunction with the accompanying drawings. The embodiments described herein are only used to more clearly illustrate the technical solutions of the present application, and are therefore only used as examples, and cannot be used to limit the scope of protection of the present application.

[0046] Reference to "embodiment" herein means that the specific features, structures or characteristics described in conjunction with the embodiment may be included in at least one embodiment of the present application. The term "embodiment" appearing in various places in the specification does not necessarily refer to the same embodiment, nor does it particularly limit its independence or association with other embodiments. In principle, in the present application, as long as there is no technical contradiction or conflict, the various technical features mentioned in the embodiments can be combined in any way to form a corresponding implementable technical solution.

[0047] Unless otherwise defined, the technical terms used in this document have the same meanings as those generally understood by those skilled in the art to which this application belongs; the use of relevant terms in this document is only for describing specific embodiments and is not intended to limit this application.

[0048] Example 1

[0049] Please refer to Figure 1 , This embodiment provides a power line data acquisition device, including: an acquisition module 1, a processing module 2, an analysis module 3, a human-computer interaction module 4 and a verification module 5;

[0050] The acquisition module 1 is placed on the top of the shell, the processing module 2 , the analysis module 3 and the verification module 5 are placed inside the shell, and the human-computer interaction module 4 is fixed to the right side of the acquisition device, in the display shell 14 .

[0051] like Figure 3 and Figure 4As shown, the main part of the device is a shell 7, with a pan-tilt base plate 13 on the top of the shell, and a collection module 1 is installed on the upper end of the pan-tilt base plate 13. A lighting lamp 6 is darkly turned at the bottom of the shell 7 for supplementary light; an operation button 8 is arranged on the outside of the human-computer interaction module 4. The bottom of the collection module 1 has left and right locking knobs 9, left and right knobs 10, front and back locking knobs 11, and front and back knobs 12. The left and right and front and back locking knobs and knobs are combined together. The left and right locking knobs 9 are used to lock the rotation angle of the receiver in the horizontal direction. When the left and right angles are adjusted in place, locking the knobs can avoid deviation during operation. The left and right knobs 10 are used to fine-tune the horizontal angle of the collection module 1 so that it can accurately align with the path or direction of the target cable. In the front and back directions, the front and back locking knobs 11 are used to adjust the pitch angle of the collection module 1. After the front and back angles are set, the locking knobs can ensure that the device remains stable and will not tilt during operation. The front and rear knobs 12 provide a fine-tuning function, so that the acquisition module 1 can make detailed adjustments to the pitch angle according to the terrain or detection requirements, and the device is always aligned with the optimal angle of the target cable.

[0052] Collection module 1 is used to collect power line related data in real time, including pole model, line length, cable width, distance from the ground, houses, trees and phase sequence information;

[0053] Processing module 2, used for preliminary processing and sorting of the information output by acquisition module 1;

[0054] Analysis module 3 uses a preset deep learning model to analyze the processed data, identify potential risks and provide decision support;

[0055] The human-computer interaction module 4 is used to provide a friendly interface for operators, display analysis results in real time, and support user interaction;

[0056] The verification module 5 is used to control the entry of incremental data and prevent the addition of abnormal data.

[0057] The acquisition module 1 includes: a laser rangefinder and a camera module; the laser rangefinder is used to collect information about the type of electric pole, line length, distance to the ground when the line crosses or is close to ground roads, houses, trees, etc., proximity distance data, and cable width information; the camera module is used to collect images of power lines to be identified to obtain phase sequence information. In this embodiment, the laser rangefinder is arranged on a pan-tilt head that can rotate 360°.

[0058] The processing module 2 performs data preprocessing and edge computing on the information output by the acquisition module 1; data preprocessing: including data cleaning, denoising, and standardization; edge computing performs real-time processing and integration of data.

[0059] In this embodiment, data preprocessing is used to ensure the quality of the collected data and provide high-quality input for the subsequent analysis module 3. In this embodiment, according to the different information output by the acquisition module 1, that is, the different outputs of the laser rangefinder and the camera module, the data output by the laser rangefinder are: pole model, line length, distance to the ground when the line crosses or is close to ground roads, houses, trees, etc., close distance data, cable width; the output signals are all distance data, and each data is processed accordingly; the camera module is used to collect images to be identified on the power line, mainly cable images; data preprocessing can be divided into two parts: signal processing of the laser rangefinder output and signal processing of the camera module output.

[0060] Signal processing of the laser rangefinder output, data cleaning: Outlier detection: Identify and remove the distance measurement values ​​that are beyond the reasonable range. It should be understood that the removed values ​​are negative values ​​or abnormally large values ​​to eliminate erroneous measurements caused by equipment failure or environmental interference. Standardization processing: In order to make the distance measurement data suitable for subsequent analysis, the data needs to be standardized. It should be understood that data standardization is a common method. In this embodiment, normalization processing is used to convert the distance measurement data into a value between 0 and 1. Data formatting: The processed data is organized into a standard structure. In this embodiment, JSON is used to facilitate the reading and processing of the subsequent analysis module 3.

[0061] Signal processing of the camera module output, grayscale processing of the image, converting color images into grayscale images to reduce data dimensions, binarization processing, to facilitate subsequent feature extraction and analysis. Image cropping and scaling: Crop and scale images to a uniform size, such as 224x224 pixels, to ensure image quality and consistency input to the deep learning model. Feature extraction: Extract feature information from the image by using image processing algorithms such as edge detection and contour extraction.

[0062] Edge computing is mainly responsible for preliminary processing at the data collection site to reduce data transmission delays and improve system response speed. The specific functions are as follows: Real-time data processing: Realize real-time processing of data output by the laser rangefinder and camera module to ensure timely response to environmental changes and hidden danger identification. Data integration: Integrate the output data of the laser rangefinder and camera module to form a unified data structure for subsequent analysis.

[0063] The analysis module 3 includes: a data input layer, a deep learning model, a feature fusion layer, a hidden danger identification layer, and a decision support layer; the data input layer is used to receive the integrated data output by the processing module 2 and convert it into the input format required by the deep learning model;

[0064] Deep learning models, including convolutional neural networks (CNNs) for image feature extraction and long short-term memory (LSTMs) for time series data analysis;

[0065] The feature fusion layer combines the features from the laser rangefinder and the camera module, forms a comprehensive feature vector through feature engineering technology, and improves the weight of key features through the attention mechanism;

[0066] The hidden danger identification layer classifies and regresses the fused features to identify potential safety hazards, such as the small relative distance between the line and surrounding buildings and vegetation, to achieve early hidden danger discovery;

[0067] The decision support layer generates decision suggestions based on the recognition results, such as repair plans, maintenance recommendations, and fault reports.

[0068] The verification module 5 includes: a data receiving layer and an incremental data verification layer; the data receiving layer is responsible for receiving the integrated data from the processing module 2; the incremental data verification layer verifies the new data according to the set rules and thresholds, and prevents the addition of abnormal data, verifies whether the ranging data and other parameters are within the preset valid range (such as the distance cannot be a negative value and must be within a reasonable physical range), and logical checks: ensure that the logical relationship between the data is reasonable, such as the correlation between the pole model and the line height above the ground.

[0069] Time series check: Check the time series data to ensure that the order of the data is reasonable and avoid duplication or confusion of timestamps.

[0070] Anomaly Detection Algorithms:

[0071] Apply algorithms such as Z-score, IQR, etc. to perform outlier detection to further identify and mark data points that do not meet expectations.

[0072] Data labeling and classification:

[0073] Mark new data as "compliant", "pending review", or "abnormal" for staff to process later.

[0074] In this embodiment, the analysis module 3 is responsible for performing in-depth analysis on the processed data using a deep learning model. The following is a detailed description of the analysis module 3: The analysis module 3 can be divided into the following submodules:

[0075] Data input layer: Receives integrated data from processing module 2, including the distance information from the laser rangefinder and the image features of the camera module. Converts the data into the input format required by the deep learning model.

[0076] Deep learning model: Model selection: In this embodiment, a pre-trained convolutional neural network (CNN) is used to process image data and identify features and patterns in images. Long short-term memory network

[0077] LSTM: used to process time series data and analyze the changing trend of ranging information.

[0078] Feature fusion layer: The features from the laser rangefinder and the camera module are fused, and a comprehensive feature vector is formed by splicing and weighting. The weight of important features is increased through the attention mechanism, so that the model can better focus on potential hidden danger information.

[0079] Hidden danger identification layer: Classify and regress the fused features to identify potential hidden dangers, such as when the relative distance between the line and surrounding buildings and vegetation is small, so as to achieve early detection of hidden dangers.

[0080] Decision support layer: Generate corresponding decision suggestions based on the analysis results, provide repair suggestions and maintenance plans. Set early warning thresholds and send alerts in time when hidden dangers are identified.

[0081] The deep model training used in this embodiment includes the following steps:

[0082] Dataset construction: Use historical data and field data to build annotated training and test sets. The training set includes samples in normal and hidden danger states. Model training: Use common deep learning frameworks such as TensorFlow to build and train models. Use cross-validation, early stopping, and hyperparameter tuning techniques to ensure the generalization ability of the model.

[0083] Performance evaluation: Evaluate the model through indicators such as precision, recall, and F1 score to ensure the reliability of the model in practical applications. Conduct independent evaluations for different types of hidden dangers to identify the performance of the model in specific situations.

[0084] Example 2

[0085] Please refer to Figure 2 Based on Example 1, this embodiment proposes a method for collecting power line data, comprising the following steps:

[0086] S1. Data collection: The laser rangefinder and the camera module run simultaneously to collect information related to the power line; the information related to the power line includes: pole model, line length, cable width, distance from the ground, houses, trees and phase sequence information; the laser rangefinder collects: pole model, line length, cable width, distance from the ground, houses, trees; the camera module collects the image to be identified of the power line to obtain the phase sequence information;

[0087] In step S1, the laser rangefinder distance measurement algorithm requires the following inputs:

[0088] The laser distance measurement module measures two distances: LA: the distance from point C to point A; LB: the distance from point C to point B; the angle θ measured by the angle sensor is the angle between LA and LB at point C; according to the cosine theorem, if the two sides and the angle of the triangle are known, the length of the third side LAB can be solved. The calculation formula is:

[0089] S2, data preprocessing; in processing module 2, the data output by the laser rangefinder is cleaned to remove outliers; the image obtained by the camera module is subjected to image denoising and enhancement processing, the brightness, contrast and saturation of the image are adjusted, and the image is cropped and scaled to a uniform size;

[0090] S3, edge computing: process the data output by the laser rangefinder and camera module, integrate them into a unified format, reduce data transmission delay, and compress data to improve transmission efficiency;

[0091] S4, data analysis: using the deep learning model preset in the analysis module 3 to analyze the processed data, identify potential risks and send alarms in time;

[0092] Preferably, step S4 in this embodiment includes the following steps:

[0093] S4.1, data input, receiving the integrated data output by the processing module 2 through the data input layer, and converting it into the input format required by the deep learning model;

[0094] S4.2. Model selection and training. The deep learning framework TensorFlow is used for model building and training, including the use of cross-validation and hyperparameter tuning.

[0095] S4.3, feature extraction and fusion, use the trained deep learning model to extract features from the data, and analyze it by combining image features and ranging information; in the feature fusion layer, the features from the laser rangefinder and the camera module are fused, and feature engineering technology is used to form a comprehensive feature vector, and the weight of key features is increased through the attention mechanism;

[0096] S4.4. Hazard analysis: At the hazard identification layer, classification and regression analysis are performed on the fusion features to identify potential safety hazards, identify the power line image to be identified, the phase sequence, and the number of lines, and output the voltage level of the conductor accordingly. The three-phase line is a 10kV line, the four-phase line is a 0.4kV line, and the two-phase line is a 0.22kV line.

[0097] In step S4.3, a convolutional neural network is used to process image data and identify features in the image; a long short-term memory network is used to process time series data, analyze the changing trend of ranging information, and combine the attention mechanism to enhance the weight of important features.

[0098] S5, decision support: Generate decision suggestions based on the analysis results, provide maintenance suggestions, maintenance plans and fault reports to operators in real time through the human-computer interaction module 4, and set early warning thresholds so that hidden dangers can be identified;

[0099] S6, data verification; in the verification module 5, the incremental data is verified to ensure that only compliant data is entered into the system and to prevent the addition of abnormal data.

[0100] The above are only preferred specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. The substitution may be a substitution of a part of the structure, device, method step, or a complete technical solution. The equivalent replacement or change according to the technical solution and the inventive concept of the present invention shall be covered within the protection scope of the present invention.

Claims

1. A power line data acquisition device, characterized in that: include: Acquisition module, processing module, analysis module, human-computer interaction module and verification module; The acquisition module is used to collect power line related data in real time, including pole model, line length, cable width, distance from the ground, houses, trees and phase sequence information; The processing module is used to preliminarily process and organize the information output by the acquisition module; The analysis module uses a preset deep learning model to analyze the processed data, identify potential risks and provide decision support; The human-computer interaction module is used to provide a friendly interface for operators, display analysis results in real time, and conduct user interaction; The verification module is used to control the entry of incremental data and prevent the addition of abnormal data.

2. The power line data acquisition device according to claim 1, characterized in that: The acquisition module includes: a laser rangefinder and a camera module; the laser rangefinder is used to collect the pole model, line length, distance to the ground when the line crosses or is close to ground roads, houses, trees, etc., proximity distance data, and cable width information; the camera module is used to collect images of power lines to be identified.

3. The power line data acquisition device according to claim 1, characterized in that: The processing module performs data preprocessing and edge computing on the information output by the acquisition module; the data preprocessing includes data cleaning, denoising, and standardization; the edge computing performs real-time processing and integration of data.

4. The power line data acquisition device according to claim 1, characterized in that: The analysis module includes: a data input layer, a deep learning model, a feature fusion layer, a hidden danger identification layer and a decision support layer; the data input layer is used to receive the integrated data output by the processing module and convert it into the input format required by the deep learning model; The deep learning model includes a convolutional neural network (CNN) for image feature extraction and a long short-term memory (LSTM) network for time series data analysis; The feature fusion layer fuses the features from the laser rangefinder and the camera module, forms a comprehensive feature vector through feature engineering technology, and improves the weight of key features through the attention mechanism; The hidden danger identification layer classifies and regresses the fused features to identify potential safety hazards; the decision support layer generates decision suggestions based on the identification results.

5. The power line data acquisition device according to claim 3, characterized in that: The verification module includes: a data receiving layer and an incremental data verification layer; the data receiving layer is responsible for receiving the integrated data from the processing module; The incremental data verification layer verifies new data according to set rules and thresholds, and prevents the addition of abnormal data.

6. A method for collecting power line data, using the power line data collection device according to any one of claims 1 to 5, characterized in that: The steps include: S1. Data collection: The laser rangefinder and the camera module run simultaneously to collect information related to the power lines; the information related to the power lines includes: pole model, line length, cable width, distance from the ground, houses, trees, and phase sequence information; S2, data preprocessing: In the processing module, the data output by the laser rangefinder is cleaned to remove outliers; the image obtained by the camera module is subjected to image denoising and enhancement processing, the brightness, contrast and saturation of the image are adjusted, and the image is cropped and scaled to a uniform size; S3, edge computing: process the data output by the laser rangefinder and camera module, integrate them into a unified format, reduce data transmission delay, and compress data to improve transmission efficiency; S4. Data analysis: Use the deep learning model preset in the analysis module to analyze the processed data, identify potential risks and send alarms in time; S5. Decision support: Generate decision suggestions based on the analysis results, provide maintenance suggestions, maintenance plans and fault reports to operators in real time through the human-computer interaction module, and set early warning thresholds so that hidden dangers can be identified; S6. Data verification: In the verification module, the incremental data is verified to ensure that only compliant data is entered into the system and to prevent the addition of abnormal data.

7. The power line data collection method according to claim 6, characterized in that: The step S4 includes the following steps: S4.1, data input, receiving the integrated data output by the processing module through the data input layer and converting it into the input format required by the deep learning model; S4.

2. Model selection and training. The deep learning framework TensorFlow is used for model building and training, including the use of cross-validation and hyperparameter tuning. S4.3, feature extraction and fusion, use the trained deep learning model to extract features from the data, and analyze it by combining image features and ranging information; in the feature fusion layer, the features from the laser rangefinder and the camera module are fused, and feature engineering technology is used to form a comprehensive feature vector, and the weight of key features is increased through the attention mechanism; S4.

4. Hazard analysis: At the hazard identification layer, classification and regression analysis are performed on the fusion features to identify potential safety hazards.

8. The power line data collection method according to claim 7, characterized in that: In step S4.3, a convolutional neural network is used to process image data and identify features in the image; a long short-term memory network or a recurrent neural network is used to process time series data, analyze the changing trend of ranging information, and combine the attention mechanism to enhance the weight of important features.

9. The power line data collection method according to claim 6, characterized in that: In step S1, the laser rangefinder needs the following inputs for distance measurement: two distances measured by the laser rangefinder module: LA: the distance from point C to point A, LB: the distance from point C to point B; and the angle θ measured by the angle sensor, i.e., the angle between LA and LB at point C.

Images