Power equipment detection method and system based on image recognition

Abstract

The present invention relates to the technical field of power equipment operation and maintenance, and in particular to a power equipment detection method and system based on image recognition, which not only improves the efficiency and accuracy of detection, but also enhances the comprehensiveness of monitoring and the ability of early fault diagnosis, and is conducive to achieving safe and stable operation of the power system; the method comprises: using a drone to collect the device appearance image and the device infrared image of the power equipment in real time; using a pre-built power equipment recognition model to perform feature recognition on the device appearance image to obtain the type of the power equipment; using the power equipment type as a filtering condition, traversing a pre-built device temperature database to obtain a focus position set and a focus position threshold temperature set corresponding to the power equipment type; based on the focus position set, extracting the temperature value of the corresponding position on the device infrared image to generate a focus position real-time temperature set.

Description

Technical Field

[0001] The present invention relates to the technical field of power equipment operation and maintenance, and in particular to a power equipment detection method and system based on image recognition. Background Art

[0002] The safe and stable operation of the power system is crucial to ensuring social and economic activities. Power equipment, such as transformers, circuit breakers, insulators, etc., may have defects or failures during long-term operation. If these defects are not detected and maintained in time, they may cause power accidents and affect power supply reliability. Therefore, regular inspection and status monitoring of power equipment is a key link in ensuring the safety of the power system.

[0003] Traditional power equipment detection methods rely on manual inspections or simple electrical parameter sensor threshold alarms. Inspectors need to carry professional equipment and go to the site to check the power equipment one by one. Not only is the workload heavy, but the inspection work is extremely difficult under complex terrain or severe weather conditions, which can easily lead to blind spots or omissions in detection. At the same time, manual inspections are affected by the inspectors' experience, skills and subjective judgment, and the detection results are uncertain. Simple threshold alarms cannot diagnose problems with power equipment in a timely manner. Therefore, a power equipment detection method based on image recognition is urgently needed. Summary of the invention

[0004] In order to solve the above technical problems, the present invention provides an image recognition-based power equipment detection method and system, which not only improves the efficiency and accuracy of detection, but also enhances the comprehensiveness of monitoring and early fault diagnosis capabilities, and helps to achieve safe and stable operation of the power system.

[0005] In a first aspect, the present invention provides a method for detecting electric power equipment based on image recognition, the method comprising:

[0006] Use drones to collect real-time appearance images and infrared images of power equipment;

[0007] Use the pre-built power equipment recognition model to perform feature recognition on the equipment appearance image to obtain the type of power equipment;

[0008] Using the power equipment type as a filtering condition, traverse the pre-built equipment temperature database to obtain a focus position set and a focus position threshold temperature set corresponding to the power equipment type;

[0009] Based on the focus position set, extracting the temperature value of the corresponding position on the infrared image of the device to generate a focus position real-time temperature set;

[0010] Perform over-temperature analysis on the focus position real-time temperature set and the focus position temperature threshold set to obtain the over-temperature state vector of the power equipment;

[0011] The over-temperature state vector of the power equipment is input into the power equipment state analysis model to obtain the real-time fault type of the power equipment.

[0012] Furthermore, the method for collecting the appearance image and infrared image of the power equipment in real time by using the drone includes:

[0013] Plan the flight route of the UAV in the ground control station according to the distribution of power equipment and terrain characteristics;

[0014] Start the drone, take off and cruise according to the planned flight route; during the cruise, the drone automatically takes appearance images and infrared images of the power equipment according to the preset photography parameters;

[0015] Transmitting image data to a ground control station via a wireless communication link;

[0016] Monitor the status of the drone and changes in the surrounding environment;

[0017] The received appearance image and infrared image of the electric power equipment are stored in the ground control station.

[0018] Furthermore, the method for constructing the device temperature database includes:

[0019] Define the data structure of the database, including a power equipment type table, a focus position table, and a temperature threshold table;

[0020] Correlating the table of power equipment types, the table of focal positions, and the table of temperature thresholds;

[0021] Collect technical manuals, maintenance guides, and industry standards for various types of power equipment to extract information about the equipment's focal locations and their temperature thresholds;

[0022] Clean the collected data to remove duplicate, erroneous and inconsistent information;

[0023] Analyze the equipment's historical operating data, fault records, and temperature monitoring data to verify and supplement information;

[0024] Classify and encode device type and focus position information;

[0025] Enter the cleaned and organized data into the database.

[0026] Furthermore, the method for acquiring the focus position real-time temperature set includes:

[0027] Calibrate the infrared image of the device, including adjusting the brightness, contrast and radiation correction of the image;

[0028] Perform image registration on the device appearance image and the infrared image so that the same device position in the two images can be aligned;

[0029] Based on the focus position set, each focus position is determined on the infrared image; and the corresponding temperature abnormality area in the device infrared image is identified;

[0030] According to the focal position, the temperature data of the pixel point corresponding to the focal position is read;

[0031] Correct and smooth the read temperature data;

[0032] The temperature values ​​corresponding to each focal position are sorted into a real-time temperature set.

[0033] Furthermore, the method for performing over-temperature analysis on the focus position real-time temperature set and the focus position temperature threshold set includes:

[0034] For each temperature value in the focus position real-time temperature set, find the temperature threshold of the corresponding focus position in the focus position temperature threshold set for comparison; the comparison results include the real-time temperature being within the normal temperature range, the real-time temperature exceeding the warning temperature threshold but not reaching the over-temperature threshold, and the real-time temperature exceeding the over-temperature threshold;

[0035] For the comparison result of each focal position, record its temperature state, wherein the temperature state includes a normal state, a warning state and an over-temperature state;

[0036] Arrange the states of all focal positions in order to form an over-temperature state vector.

[0037] Furthermore, when performing secondary inspection on multiple power equipment in the inspection area where no fault is found, the method for setting the inspection route of the drone includes:

[0038] Collecting power equipment over-temperature state vectors of a plurality of power equipments for which no fault is found;

[0039] For each power equipment over-temperature state vector of each power equipment that has not been found to have a fault, an over-temperature statistical analysis is performed to obtain the over-temperature state index of the power equipment; the higher the over-temperature state index, the more likely the power equipment is to fail;

[0040] Based on the over-temperature status index of the power equipment, multiple power equipment with no faults found are sorted from high to low to obtain a power equipment status sequence;

[0041] The detection route of the drone is generated according to the status sequence of the power equipment. The power equipment that is closer to the front in the status sequence of the power equipment will be detected first.

[0042] Furthermore, the structure of the power equipment identification model includes:

[0043] Input layer: used to receive standardized image data;

[0044] Convolution layer: The input image is convolved with a learnable convolution kernel to extract local features from the image.

[0045] Region Proposal Network: used to generate candidate regions that contain power equipment;

[0046] Pooling layer: Perform pooling operation on each candidate area to obtain a feature map of fixed size;

[0047] Fully connected layer: pass the feature map to the fully connected layer for classification and regression;

[0048] Classification branch: used to predict the category of each candidate region;

[0049] Regression branch: used to fine-tune the location of each candidate region.

[0050] On the other hand, the present application also provides an electric power equipment detection system based on image recognition, the system comprising:

[0051] Image acquisition module, using drones to collect real-time appearance images and infrared images of power equipment;

[0052] The device identification module uses the pre-built power device identification model to perform feature recognition on the device appearance image to obtain the type of power device;

[0053] The temperature database query module uses the power equipment type as a filtering condition, traverses the pre-built equipment temperature database, and obtains a focus position set and a focus position threshold temperature set corresponding to the power equipment type; wherein the focus position set includes all focus positions that need to be focused on for the power equipment type, and the focus position temperature threshold set includes the temperature threshold corresponding to each focus position;

[0054] The infrared image temperature extraction module extracts the temperature value of the corresponding position on the infrared image of the device based on the focus position set, and generates a real-time temperature set of the focus position;

[0055] An over-temperature analysis module performs over-temperature analysis on the focus position real-time temperature set and the focus position temperature threshold set to obtain an over-temperature state vector of the power equipment;

[0056] The fault analysis module inputs the over-temperature state vector of the power equipment into the power equipment state analysis model to obtain the real-time fault type of the power equipment.

[0057] In a third aspect, the present application provides an electronic device, comprising a bus, a transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the transceiver, the memory, and the processor are connected via the bus, and the computer program, when executed by the processor, implements the steps of any one of the above methods.

[0058] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the steps in any one of the above-mentioned methods when executed by a processor.

[0059] Compared with the prior art, the invention has the following beneficial effects: by collecting images of power equipment in real time through drones, the appearance and temperature information of the equipment can be quickly obtained, which greatly improves the detection efficiency; at the same time, by using computer vision and image recognition technology to process image data in real time, the status information of the power equipment can be quickly obtained, which ensures the real-time nature of the detection;

[0060] This method not only focuses on the appearance characteristics of power equipment, but also monitors the temperature distribution of the equipment through infrared images, which can comprehensively capture the operating status of the equipment. By using the pre-built power equipment recognition model and temperature database, it can accurately identify the equipment type and match the corresponding focus position and temperature threshold, thus improving the accuracy of detection.

[0061] Compared with manual inspection, this method reduces the impact of inspectors’ experience, skills and subjective judgment on the inspection results; computer vision and image recognition technology have higher objectivity and consistency, which can reduce human errors and improve the reliability of inspection results;

[0062] Drones can collect images in complex terrain or bad weather conditions, overcoming the limitations of manual inspections. At the same time, this method can also be customized according to the characteristics and needs of different power equipment, with strong adaptability and flexibility.

[0063] By analyzing the temperature information of the infrared image of the equipment in real time, the over-temperature state of the equipment can be discovered in time, providing early warning information for equipment maintenance; at the same time, the over-temperature state vector is input into the power equipment state analysis model, which can further diagnose the real-time fault type of the equipment and provide a basis for maintenance decisions;

[0064] The entire detection process is automated and intelligent, reducing manual intervention and improving work efficiency. By continuously optimizing and upgrading the recognition model, temperature database and state analysis model, the performance and accuracy of the detection method can be further improved.

[0065] In summary, the power equipment detection method based on image recognition not only improves the efficiency and accuracy of detection, but also enhances the comprehensiveness of monitoring and early fault diagnosis capabilities, which helps to achieve safe and stable operation of the power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 is a flow chart of the present invention;

[0067] Figure 2 It is a flow chart for collecting the equipment appearance image and equipment infrared image of the power equipment;

[0068] Figure 3 It is a structural diagram of the power equipment detection system based on image recognition. DETAILED DESCRIPTION

[0069] In the description of the present application, those skilled in the art should know that the present application can be implemented as a method, an apparatus, an electronic device, and a computer-readable storage medium. Therefore, the present application can be specifically implemented in the following forms: complete hardware, complete software (including firmware, resident software, microcode, etc.), a combination of hardware and software. In addition, in some embodiments, the present application can also be implemented in the form of a computer program product in one or more computer-readable storage media, and the computer-readable storage medium contains computer program code.

[0070] The above-mentioned computer-readable storage medium can adopt any combination of one or more computer-readable storage media. Computer-readable storage media include: electrical, magnetic, optical, electromagnetic, infrared or semiconductor systems, devices or devices, or any combination of the above. More specific examples of computer-readable storage media include: portable computer disks, hard disks, random access memories, read-only memories, erasable programmable read-only memories, flash memories, optical fibers, optical disc read-only memories, optical storage devices, magnetic storage devices, or any combination of the above. In the present application, computer-readable storage media can be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device, or device.

[0071] The acquisition, storage, use, and processing of data in the technical solution of this application comply with the relevant provisions of national laws.

[0072] The present application describes the provided methods, devices, and electronic devices through flowcharts and / or block diagrams.

[0073] It should be understood that each box in the flowchart and / or block diagram and the combination of boxes in the flowchart and / or block diagram can be implemented by computer-readable program instructions. These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer or other programmable data processing device to produce a machine, and these computer-readable program instructions are executed by a computer or other programmable data processing device to produce a device that implements the functions / operations specified by the boxes in the flowchart and / or block diagram.

[0074] These computer-readable program instructions can also be stored in a computer-readable storage medium that enables a computer or other programmable data processing device to work in a specific manner. In this way, the instructions stored in the computer-readable storage medium produce an instruction device product including functions / operations specified in the blocks in the flowchart and / or block diagram.

[0075] Computer-readable program instructions can also be loaded onto a computer, other programmable data processing apparatus, or other device, so that a series of operating steps are performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable data processing apparatus can provide a process for implementing the functions / operations specified in the blocks in the flowchart and / or block diagram.

[0076] The present application is described below in conjunction with the drawings in the present application.

[0077] Embodiment 1: Figure 1 to Figure 2 As shown, the power equipment detection method based on image recognition of the present invention specifically includes the following steps:

[0078] S1. Use drones to collect real-time appearance images and infrared images of power equipment;

[0079] Select the drone model based on the distribution characteristics of power equipment, the operating environment, and safety requirements; the drone should have sufficient load capacity to carry high-definition cameras and infrared thermal imagers, and also have good flight stability and wind resistance;

[0080] Install high-resolution visible light cameras and infrared thermal imagers on drones; visible light cameras are used to capture images of the appearance of power equipment to identify the type and status of the equipment; infrared thermal imagers are used to detect the temperature distribution of the equipment and identify potential overheating problems;

[0081] Ensure stable and reliable communication between the UAV and the ground control station to transmit image data and receive control instructions in real time; the UAV should be equipped with a GPS navigation system to ensure accurate flight and positioning;

[0082] The method for collecting the appearance image and infrared image of the power equipment in real time by using the unmanned aerial vehicle comprises:

[0083] Plan the flight route of the drone in the ground control station according to the distribution of power equipment and terrain characteristics; the route should cover all power equipment to be inspected and ensure that the appearance and infrared image of the equipment can be clearly captured during the flight;

[0084] Start the drone, take off and cruise according to the planned flight route; during the cruise, the drone will automatically take appearance images and infrared images of the power equipment according to the preset photography parameters;

[0085] The image stabilization system on the drone will ensure that the captured images are clear and stable; the image data will be transmitted to the ground control station via a wireless communication link in real time for subsequent processing and analysis;

[0086] During the flight of the drone, the ground control station monitors the status of the drone and changes in the surrounding environment in real time; once an abnormal situation is found, emergency measures will be taken immediately to ensure the safety of the drone and personnel;

[0087] The received appearance images and infrared images of the power equipment shall be stored in the ground control station; the image data shall be classified and indexed according to the equipment type, shooting time, location and other information for subsequent retrieval and analysis.

[0088] In this step, high-altitude inspections by drones can quickly cover a wide range of power equipment areas, greatly reducing the time and labor costs required for manual inspections; the high-speed flight and automated shooting functions of drones significantly improve detection efficiency; the high-resolution visible light cameras and infrared thermal imagers carried by drones can capture the subtle appearance features and precise temperature distribution information of power equipment; they help to more accurately identify the type and status of equipment, and promptly detect potential faults such as overheating; manual inspections in complex terrain or severe weather conditions pose a greater safety risk; drone inspections can prevent inspectors from being directly exposed to dangerous environments, reducing the probability of safety accidents; drones It is able to accurately fly and position according to the preset flight route, ensuring comprehensive coverage of all power equipment to be inspected; this helps to reduce detection blind spots or omissions caused by the difficulty of manual inspections; the received appearance images and infrared images of the power equipment are stored in the ground control station, and classified and indexed according to equipment type, shooting time, location and other information, which facilitates subsequent rapid retrieval and analysis of image data; this provides strong data support for the operation and maintenance management of power equipment; this step uses drones to collect equipment appearance images and equipment infrared images of power equipment in real time, which not only improves detection efficiency and accuracy, reduces safety risks, but also reduces detection blind spots and improves data quality and management efficiency.

[0089] S2. Using a pre-built power equipment recognition model to perform feature recognition on the equipment appearance image to obtain the type of power equipment;

[0090] Preprocess the appearance images of power equipment collected by drones in real time, including resizing, normalization, and denoising, to improve image quality and model recognition effect;

[0091] The preprocessed image is input into the power equipment recognition model; the model performs convolution operations on the image layer by layer to extract different levels of features from low-level to high-level;

[0092] Based on feature extraction, the model performs weighted summation and activation operations on these features, and finally outputs a probability distribution vector, which indicates the probability that the input image belongs to each type of power equipment. By comparing the size of these probability values, the type of power equipment in the input image can be determined.

[0093] The type of power equipment identified by the model is used as the output result for subsequent equipment temperature database query and status analysis;

[0094] The structure of the electric power equipment identification model includes:

[0095] Input layer: used to receive standardized image data;

[0096] Convolutional layer: Multiple convolutional layers are stacked, each followed by an activation function, and the input image is convolved with a learnable convolution kernel to extract local features in the image;

[0097] Region Proposal Network: used to generate candidate regions that contain power equipment;

[0098] Pooling layer: Perform pooling operation on each candidate area to obtain a feature map of fixed size;

[0099] Fully connected layer: pass the feature map to the fully connected layer for classification and regression;

[0100] Classification branch: used to predict the category of each candidate region;

[0101] Regression branch: used to fine-tune the location of each candidate region.

[0102] In this step, the model performs layer-by-layer convolution operations on the image, which can effectively extract different levels of features from low-level to high-level, which helps to improve the accuracy of recognition; at the same time, the weighted summation and activation operations within the model further enhance the expressiveness of the features, making the final output probability distribution vector more reliable; image preprocessing can significantly improve image quality and reduce the impact of noise and interference on the recognition results, thereby enhancing the robustness of the model; this enables the model to maintain good recognition effects under different lighting conditions, shooting angles and equipment states; compared with traditional manual inspection methods, the power equipment detection method based on image recognition has achieved automation and greatly improved work efficiency; drones can quickly collect a large amount of image data, and the model can complete the processing and recognition of these images in a short time, greatly shortening the detection cycle; the type of power equipment identified by the model as the output result can not only be directly used for fault warning and maintenance guidance, but also as an important basis for subsequent equipment temperature database query and status analysis; this helps to achieve comprehensive monitoring and refined management of the operating status of power equipment.

[0103] S3, using the power equipment type as a filtering condition, traversing the pre-built equipment temperature database to obtain a focus position set and a focus position threshold temperature set corresponding to the power equipment type; wherein the focus position set includes all focus positions that need to be focused on for the power equipment type, and the focus position temperature threshold set includes the temperature threshold corresponding to each focus position;

[0104] Use the power equipment type as a query index and initiate a query request to the pre-built equipment temperature database;

[0105] The database system retrieves a focus position set and a corresponding focus position temperature threshold set matching the type of the electric power equipment according to the query request;

[0106] Verify the retrieved focus position and temperature threshold data to ensure the accuracy, timeliness and completeness of the data;

[0107] The verified focus position set and focus position temperature threshold set are used as output results and passed to the subsequent steps for extracting the temperature value of the corresponding position on the device infrared image and performing over-temperature analysis;

[0108] The method for constructing the device temperature database comprises:

[0109] Define the data structure of the database, including the power equipment type table, the focus position table and the temperature threshold table; the power equipment type table is used to store the classification information of the equipment; the focus position table is associated with the specific equipment type and lists the focus position that needs to be monitored for each equipment; the temperature threshold table records the temperature threshold of each focus position under different working conditions;

[0110] Correlate the power equipment type table, focus position table, and temperature threshold table to ensure data integrity and consistency;

[0111] Collect technical manuals, maintenance guides, and industry standards for various types of power equipment to extract information about the equipment's focal locations and their temperature thresholds;

[0112] Clean the collected data to remove duplicate, erroneous or inconsistent information; ensure the accuracy and reliability of the data;

[0113] Analyze the historical operation data, fault records and temperature monitoring data of the equipment to verify and supplement the information in the technical documentation; historical data can provide the temperature change trend and failure mode of the equipment in actual operation, which helps to set the temperature threshold more accurately;

[0114] Classify and encode device type and focus location information for easy storage and retrieval in the database;

[0115] Entering the cleaned and sorted data into a database; including adding equipment classification information in the power equipment type table, listing the focal position of each device in the focal position table, and setting the temperature threshold of each focal position in the temperature threshold table;

[0116] In this step, by using the power equipment type as a query index, the focus position set and temperature threshold set of a specific equipment type can be quickly located; the efficiency of data retrieval is greatly improved, while also ensuring the accuracy of the query results; it is helpful to accurately extract the temperature value on the equipment infrared image in subsequent steps and conduct effective over-temperature analysis; after retrieving the focus position and temperature threshold data, a strict verification process is carried out to ensure the accuracy, timeliness and integrity of the data; it effectively avoids misjudgments caused by incorrect or outdated data, and further improves the reliability of power equipment detection; the equipment temperature database realizes the automated management and intelligent application of data; this not only reduces the burden of manual inspections, but also improves the intelligence level of detection work; this step provides strong technical support for power equipment detection through efficient data retrieval, strict data verification and intelligent data application, significantly improves the efficiency and accuracy of detection work, and provides important guarantees for the safe and stable operation of the power system.

[0117] S4, based on the focus position set, extracting the temperature value of the corresponding position on the infrared image of the device to generate a focus position real-time temperature set;

[0118] The method for acquiring the focus position real-time temperature set includes:

[0119] Calibrate the infrared images collected by the drone to ensure that the temperature information in the image is accurate; this includes adjusting the brightness and contrast of the image and performing necessary radiation correction to eliminate the impact of environmental factors on the infrared image;

[0120] Through feature point matching and image transformation technology, the device appearance image and the infrared image are registered to ensure that the same device position in the two images can be accurately aligned;

[0121] Based on the focus position set, use image processing technology to accurately find each focus position on the infrared image; help identify the corresponding hot spots or temperature abnormality areas in the device infrared image;

[0122] In an infrared image, each pixel corresponds to a temperature value; based on the precise positioning of the focus position, the temperature data of the pixels corresponding to these positions are read;

[0123] Further correction and smoothing of the read temperature data helps to improve the accuracy and stability of the temperature data;

[0124] The temperature values ​​corresponding to each focal position are sorted into a real-time temperature set; this set will be used for subsequent over-temperature analysis.

[0125] In this step, the influence of environmental factors on the infrared image is effectively eliminated by carefully calibrating the infrared image of the equipment collected by the drone, thereby improving the accuracy of the temperature information in the image; the equipment appearance image and the infrared image are accurately registered by using feature point matching and image transformation technology; the accurate alignment of the same equipment position in the two images is ensured, which provides a solid foundation for the subsequent accurate positioning of the focus position on the infrared image; based on the focus position set, each focus position is accurately found on the infrared image through image processing technology, and the corresponding hot spot or temperature abnormality area is successfully identified, making the temperature monitoring of key parts of the power equipment more accurate and reliable; the read focus position temperature data is further corrected and smoothed, effectively reducing the influence of noise and error on the data, and improving the stability and reliability of the data; by sorting out the temperature values ​​corresponding to each focus position, a real-time temperature set is formed; this set not only contains rich temperature information, but also has high accuracy and stability, which provides strong data support for subsequent over-temperature analysis, power equipment status assessment and fault diagnosis; this step realizes the accurate extraction and sorting of the focus position temperature in the infrared image of the power equipment through fine operation and processing, providing important data support for the entire power equipment detection process.

[0126] S5, performing over-temperature analysis on the focus position real-time temperature set and the focus position temperature threshold set to obtain an over-temperature state vector of the power equipment;

[0127] The method for performing over-temperature analysis on a focus position real-time temperature set and a focus position temperature threshold set includes:

[0128] For each temperature value in the focus position real-time temperature set, find the temperature threshold value of the corresponding focus position in the focus position temperature threshold set for comparison;

[0129] If the real-time temperature is within the normal temperature range, the current state of the focus position is considered normal and no further processing is required;

[0130] If the real-time temperature exceeds the warning temperature threshold but does not reach the severe over-temperature threshold, the focus position is considered to be in a warning state and requires further monitoring and preventive maintenance measures;

[0131] If the real-time temperature exceeds the severe over-temperature threshold, it is considered that there is a severe over-temperature problem at the focus position, and immediate intervention measures are required to prevent equipment damage or accidents;

[0132] For the comparison result of each focal position, record its temperature status, wherein the temperature status includes a normal status, a warning status and a serious over-temperature status;

[0133] The states of all focal positions are arranged in order to form an over-temperature state vector; the over-temperature state vector intuitively reflects the over-temperature conditions of each focal position of the power equipment at present.

[0134] In this step, the temperature of each focal position of the power equipment can be monitored and compared in real time to ensure timely detection of over-temperature conditions, thereby improving the accuracy and timeliness of detection; temperature monitoring and comparison are carried out for each key focal position of the power equipment to ensure the comprehensiveness of the detection; at the same time, by setting temperature thresholds of different levels, the degree of over-temperature at each focal position can be carefully evaluated, providing a more accurate basis for subsequent maintenance decisions; for the warning state and serious over-temperature state found, the corresponding alarm mechanism can be triggered immediately to notify the operation and maintenance personnel to handle it; it is helpful to take preventive maintenance measures or emergency intervention measures in a timely manner to prevent equipment damage or accidents, and ensure the safe and stable operation of the power system; the construction of the over-temperature state vector provides the operation and maintenance personnel with intuitive and visual equipment status information; by viewing the over-temperature state vector, the operation and maintenance personnel can quickly understand the overall over-temperature situation of the power equipment, providing strong support for the formulation of maintenance plans and decisions.

[0135] S6. Inputting the over-temperature state vector of the power equipment into the power equipment state analysis model to obtain the real-time fault type of the power equipment;

[0136] Encode the over-temperature state vector of the power equipment to ensure that its format matches the requirements of the model input layer;

[0137] Once the over-temperature state vector is input into the power equipment state analysis model, the model will perform inference calculations based on the learned mapping relationship, extract useful information from the input data and predict the real-time fault type of the power equipment;

[0138] After the model reasoning is completed, the prediction results are output, indicating the real-time fault type of the power equipment;

[0139] The real-time fault types of the power equipment include:

[0140] Poor contact: The contact resistance between conductors increases, causing local heating;

[0141] Overload: The electrical equipment operates beyond its rated current or power, causing heating;

[0142] Insulation damage: The performance of the insulation material deteriorates due to aging, moisture, etc., causing leakage and heat;

[0143] Short circuit: The formation of an abnormal conductive path in the power system, causing abnormal increase in current and resulting in equipment damage;

[0144] Partial discharge: The discharge phenomenon that occurs in high-voltage equipment is accompanied by local heating;

[0145] Mechanical failure: damage to mechanical parts due to wear, corrosion, vibration, etc.;

[0146] Cooling system failure: cooling medium circulation is not smooth or cooling capacity is reduced.

[0147] In this step, through the automated image recognition and state analysis model, the fault detection of power equipment can be carried out quickly, thus improving work efficiency; the state analysis model based on image recognition and machine learning can make objective and accurate judgments based on pre-learned knowledge and rules, thus reducing errors caused by human factors; by using remote acquisition equipment such as drones, the appearance and infrared images of the power equipment can be obtained in real time without direct contact with the equipment, and then the state analysis can be carried out; this not only reduces the safety risks of inspection personnel, but also makes monitoring possible in complex terrain or severe weather conditions; the model can identify various types of faults through in-depth analysis of the over-temperature state vector, which helps to detect problems in a timely manner and take corresponding measures to prevent the fault from expanding or causing more serious consequences; by timely detecting and handling potential faults of power equipment, it can effectively reduce power outages caused by equipment failures, thereby improving the reliability and stability of power supply.

[0148] Furthermore, in the subsequent inspection operation, it is necessary to conduct secondary inspection on multiple power equipment in the inspection area that have not been found to have faults. When conducting the secondary inspection, the method for setting the UAV inspection route includes:

[0149] Step T1, collecting power equipment over-temperature state vectors of multiple power equipments that have not been found to have faults; the power equipment over-temperature state vector is obtained by performing over-temperature analysis on the infrared image of the power equipment in the previous step, reflects the real-time temperature state of the power equipment, and is an important basis for evaluating whether the equipment may fail;

[0150] Step T2: For each power equipment over-temperature state vector of the power equipment that has not been found to have a fault, an over-temperature statistical analysis is performed to obtain the over-temperature state index of the power equipment; the higher the over-temperature state index, the more likely the power equipment is to fail; including statistically analyzing the temperature state in the over-temperature state vector of each device, and calculating the over-temperature state index of each device based on the statistical results, the higher the over-temperature state index, the more likely the power equipment is to fail; the specific calculation method of the over-temperature state index is as follows: each element in the over-temperature state vector represents the temperature state of a focal position, and the over-temperature state index is quantified by assigning weights or scores to different states. temperature state; for example, the normal state is assigned the lowest score, the warning state is assigned the middle score, and the over-temperature state is assigned the highest score; the quantified results of each focal position are weighted averaged or otherwise comprehensively calculated to obtain an overall over-temperature state index; the selection of weights is determined based on the importance of the focal position or the degree of its impact on the overall operation of the power equipment; in order to facilitate comparison of the over-temperature states of different power equipment, the calculated over-temperature state index is standardized so that it falls within a fixed range, such as between 0 and 1, where 0 indicates that there is no over-temperature at all and 1 indicates that all focal positions are in a serious over-temperature state;

[0151] Step T3: Based on the over-temperature status index of the power equipment, multiple power equipments that have not been found to have a fault are sorted from high to low to obtain a power equipment status sequence; the power equipment status sequence is a priority ranking of the power equipment according to its potential fault risk;

[0152] Step T4: Generate the detection route of the drone according to the power equipment status sequence. In the power equipment status sequence, the power equipment that is closer to the front will be detected first. Plan the detection route of the drone according to the equipment status sequence, and give priority to checking those power equipment with higher over-temperature status index, that is, more prone to problems, to ensure that the equipment most likely to fail is checked first, so as to discover and solve the problem in time.

[0153] In the above process, by calculating the over-temperature status index, the power equipment with higher potential failure risks can be quickly identified, so that detection can be prioritized, unnecessary inspection time can be reduced, and overall detection efficiency can be improved; by analyzing the over-temperature status of power equipment, it can be predicted which equipment is more likely to fail, thereby achieving early warning and preventive maintenance and reducing the probability of failure; by planning the inspection route according to the status sequence of power equipment, drones and human resources can be reasonably allocated, giving priority to equipment with higher risks and achieving optimal resource allocation.

[0154] Embodiment 2: Figure 3 As shown, the power equipment detection system based on image recognition of the present invention specifically includes the following modules;

[0155] Image acquisition module, using drones to collect real-time appearance images and infrared images of power equipment;

[0156] The device identification module uses the pre-built power device identification model to perform feature recognition on the device appearance image to obtain the type of power device;

[0157] The temperature database query module uses the power equipment type as a filtering condition, traverses the pre-built equipment temperature database, and obtains a focus position set and a focus position threshold temperature set corresponding to the power equipment type; wherein the focus position set includes all focus positions that need to be focused on for the power equipment type, and the focus position temperature threshold set includes the temperature threshold corresponding to each focus position;

[0158] The infrared image temperature extraction module extracts the temperature value of the corresponding position on the infrared image of the device based on the focus position set, and generates a real-time temperature set of the focus position;

[0159] An over-temperature analysis module performs over-temperature analysis on the focus position real-time temperature set and the focus position temperature threshold set to obtain an over-temperature state vector of the power equipment;

[0160] The fault analysis module inputs the over-temperature state vector of the power equipment into the power equipment state analysis model to obtain the real-time fault type of the power equipment.

[0161] The system realizes the automation of the whole process from image acquisition to fault analysis, greatly reducing manual intervention and improving the intelligent level of detection. The image acquisition by drone avoids the tediousness and danger of manual inspection.

[0162] The system not only collects images of the equipment's appearance, but also uses infrared images for temperature detection, which can comprehensively monitor the operating status of power equipment. Through the pre-built power equipment recognition model and temperature database, the system can accurately identify the equipment type and obtain the corresponding focus position and temperature threshold, thereby performing precise temperature analysis and fault diagnosis.

[0163] The system can collect and process image data in real time, quickly generate over-temperature state vectors of power equipment, and input them into the fault analysis model to obtain real-time fault types, which can effectively avoid the occurrence of power accidents.

[0164] Compared with the traditional manual inspection method, the system reduces the impact of the inspection personnel's experience, skills and subjective judgment on the inspection results, and reduces human errors; at the same time, the system processes and analyzes data through algorithms and models, which has higher objectivity and accuracy; the system can detect the status of multiple power equipment at the same time, greatly improving the inspection efficiency; in addition, the speed of drone image acquisition is also faster than manual inspection, further shortening the inspection cycle; the system has good scalability and flexibility, and can update and optimize the inspection model, database and algorithm according to actual needs;

[0165] In summary, the power equipment detection system based on image recognition not only improves the efficiency and accuracy of detection, but also enhances the comprehensiveness of monitoring and early fault diagnosis capabilities, which helps to achieve safe and stable operation of the power system.

[0166] The various variations and specific embodiments of the power equipment detection method based on image recognition in the aforementioned embodiment 1 are also applicable to the power equipment detection system based on image recognition in this embodiment. Through the aforementioned detailed description of the power equipment detection method based on image recognition, those skilled in the art can clearly understand the implementation method of the power equipment detection system based on image recognition in this embodiment, so for the sake of brevity of the specification, it will not be described in detail here.

[0167] In addition, the present application also provides an electronic device, including a bus, a transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor. The transceiver, the memory, and the processor are respectively connected via a bus. When the computer program is executed by the processor, each process of the above-mentioned method for controlling output data is implemented, and the same technical effect can be achieved. To avoid repetition, it will not be described here.

[0168] The above is only a preferred embodiment of the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the technical principle of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A method for detecting electric power equipment based on image recognition, characterized in that: The method comprises: Based on the pre-set inspection route, drones are used to collect real-time appearance images and infrared images of power equipment; Use the pre-built power equipment recognition model to perform feature recognition on the equipment appearance image to obtain the type of power equipment; Using the power equipment type as a filtering condition, traverse the pre-built equipment temperature database to obtain a focus position set and a focus position temperature threshold set corresponding to the power equipment type; Based on the focus position set, extracting the temperature value of the corresponding position on the infrared image of the device to generate a focus position real-time temperature set; Perform over-temperature analysis on the focus position real-time temperature set and the focus position temperature threshold set to obtain the over-temperature state vector of the power equipment; Inputting the over-temperature state vector of the power equipment into the power equipment state analysis model to obtain the real-time fault type of the power equipment; The method for generating the focus position real-time temperature set comprises: Calibrate the infrared image of the device, including adjusting the brightness, contrast and radiation correction of the image; Perform image registration on the device appearance image and the infrared image so that the same device position in the two images can be aligned; Based on the focus position set, each focus position is determined on the infrared image; and the corresponding temperature abnormality area in the device infrared image is identified; According to the focal position, the temperature data of the pixel point corresponding to the focal position is read; Correct and smooth the read temperature data; Arrange the temperature values ​​corresponding to each focal position into a real-time temperature set; The method for performing over-temperature analysis on a focus position real-time temperature set and a focus position temperature threshold set includes: For each temperature value in the focus position real-time temperature set, find the temperature threshold of the corresponding focus position in the focus position temperature threshold set for comparison; the comparison results include the real-time temperature being within the normal temperature range, the real-time temperature exceeding the warning temperature threshold but not reaching the over-temperature threshold, and the real-time temperature exceeding the over-temperature threshold; For the comparison result of each focal position, record its temperature state, wherein the temperature state includes a normal state, a warning state and an over-temperature state; Arrange the states of all focal positions in order to form an over-temperature state vector; When performing secondary inspection on multiple power equipment that have not been found to have faults in the inspection area, the method for setting the inspection route of the drone includes: Collecting power equipment over-temperature state vectors of a plurality of power equipments for which no fault is found; For each power equipment over-temperature state vector of each power equipment that has not been found to have a fault, an over-temperature statistical analysis is performed to obtain the over-temperature state index of the power equipment; the higher the over-temperature state index, the more likely the power equipment is to fail; Based on the over-temperature status index of the power equipment, multiple power equipment with no faults found are sorted from high to low to obtain a power equipment status sequence; Generate the detection route of the drone according to the power equipment status sequence. The power equipment that is closer to the front in the power equipment status sequence will be detected first. The specific calculation method of the over-temperature state index is as follows: each element in the over-temperature state vector represents the temperature state of a focal position, and the over-temperature state is quantified by assigning weights to different states; the normal state is assigned the lowest score, the warning state is assigned a medium score, and the over-temperature state is assigned the highest score; the quantified results of each focal position are comprehensively calculated to obtain an overall over-temperature state index; the choice of weight is determined according to the importance of the focal position or the degree of influence of the focal position on the overall operation of the power equipment.

2. The method for detecting electric power equipment based on image recognition according to claim 1, characterized in that: The method for collecting the appearance image and infrared image of the power equipment in real time by using the unmanned aerial vehicle comprises: Plan the flight route of the UAV in the ground control station according to the distribution of power equipment and terrain characteristics; Start the drone, take off and cruise according to the planned flight route; during the cruise, the drone automatically takes appearance images and infrared images of the power equipment according to the preset photography parameters; Transmitting image data to a ground control station via a wireless communication link; Monitor the status of the drone and changes in the surrounding environment; The received appearance image and infrared image of the electric power equipment are stored in the ground control station.

3. The method for detecting electric power equipment based on image recognition according to claim 1, characterized in that: The method for constructing the device temperature database comprises: Define the data structure of the database, including a power equipment type table, a focus position table, and a temperature threshold table; Correlating the table of power equipment types, the table of focal positions, and the table of temperature thresholds; Collect technical manuals, maintenance guides, and industry standards for various types of power equipment to extract information about the equipment's focal locations and their temperature thresholds; Clean the collected data to remove duplicate, erroneous and inconsistent information; Analyze the equipment's historical operating data, fault records, and temperature monitoring data to verify and supplement information; Classify and encode device type and focus position information; Enter the cleaned and organized data into the database.

4. The method for detecting electric power equipment based on image recognition according to claim 1, characterized in that: The structure of the electric power equipment identification model includes: Input layer: used to receive standardized image data; Convolution layer: The input image is convolved with a learnable convolution kernel to extract local features from the image. Region Proposal Network: used to generate candidate regions that contain power equipment; Pooling layer: Perform pooling operation on each candidate area to obtain a feature map of fixed size; Fully connected layer: pass the feature map to the fully connected layer for classification and regression; Classification branch: used to predict the category of each candidate region; Regression branch: used to fine-tune the location of each candidate region.

5. An electric power equipment detection system based on image recognition, characterized in that: The system is applied to the power equipment detection method based on image recognition as claimed in claim 1, and the system comprises: Image acquisition module, using drones to collect real-time appearance images and infrared images of power equipment; The device identification module uses a pre-built power device identification model to perform feature recognition on the device appearance image to obtain the type of power device; The temperature database query module uses the power equipment type as a filtering condition, traverses the pre-built equipment temperature database, and obtains a focus position set and a focus position temperature threshold set corresponding to the power equipment type; wherein the focus position set includes all focus positions that need to be focused on for the power equipment type, and the focus position temperature threshold set includes the temperature threshold corresponding to each focus position; The infrared image temperature extraction module extracts the temperature value of the corresponding position on the infrared image of the device based on the focus position set, and generates a real-time temperature set of the focus position; An over-temperature analysis module performs over-temperature analysis on the focus position real-time temperature set and the focus position temperature threshold set to obtain an over-temperature state vector of the power equipment; The fault analysis module inputs the over-temperature state vector of the power equipment into the power equipment state analysis model to obtain the real-time fault type of the power equipment.

6. An electronic device for detecting power equipment based on image recognition, comprising a bus, a transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the transceiver, the memory, and the processor are connected via the bus, characterized in that: When the computer program is executed by the processor, the steps in the method according to any one of claims 1 to 4 are implemented.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps in the method according to any one of claims 1 to 4 are implemented.

Images