Power transmission and transformation equipment detection method and system based on deep learning, and terminal

Abstract

The invention provides a deep learning-based power transmission and transformation equipment detection method, system and terminal, and relates to the technical field of power transmission and transformation equipment detection, and the method comprises the steps: obtaining an initial infrared image and an equipment detection image of power transmission and transformation equipment; controlling a preset target detection model to analyze the equipment detection image to generate an image target detection parameter; correcting the image target detection parameter according to a preset image deviation parameter to generate an actual target detection parameter; marking the initial infrared image according to the actual target detection parameters to generate a marked infrared image; acquiring environment detection parameters of the power transmission and transformation equipment; correcting the labeled infrared image according to the environment detection parameters to generate an actual infrared image; and analyzing the actual infrared image to generate an equipment detection report. The method has the effect of improving the timeliness of detection of the power transmission and transformation equipment.

Description

Technical Field

[0001] This application relates to the technical field of power transmission and transformation equipment detection, and in particular to a method, system, and terminal for power transmission and transformation equipment detection based on deep learning. Background Art

[0002] Power transmission and transformation equipment detection refers to a series of detection, testing, and evaluation work on power transmission and transformation equipment in the power system to ensure the safe and reliable operation of these equipment.

[0003] In the related art, an operator carries a thermal imager and goes to each predetermined detection location one by one according to a pre-made inspection plan, uses the thermal imager to perform temperature scanning and image capture on the power transmission and transformation equipment, so as to obtain the infrared spectrum of the power transmission and transformation equipment, and then compares the infrared spectrum with the infrared image when the power transmission and transformation equipment fails to determine whether the power transmission and transformation equipment fails.

[0004] In view of the above related art, when using a thermal imager to collect the infrared spectrum of power transmission and transformation equipment, it is necessary to ensure the stability of the external environment to reduce the interference of the external environment on the accuracy of collecting the infrared spectrum. Therefore, the operator needs to select a specified time to detect the power transmission and transformation equipment, and cannot detect the power transmission and transformation equipment in real time, resulting in poor timeliness of detecting the power transmission and transformation equipment. Summary of the Invention

[0005] In order to improve the timeliness of power transmission and transformation equipment detection, this application provides a method, system, and terminal for power transmission and transformation equipment detection based on deep learning.

[0006] In a first aspect, this application provides a method for power transmission and transformation equipment detection based on deep learning, adopting the following technical solution: The method for power transmission and transformation equipment detection based on deep learning includes: Obtain the initial infrared image and equipment detection image of the power transmission and transformation equipment; Control a preset target detection model to analyze the equipment detection image to generate image target detection parameters; Correct the image target detection parameters according to the preset image deviation parameters to generate actual target detection parameters; Label the initial infrared image according to the actual target detection parameters to generate a labeled infrared image; Obtain the environmental detection parameters of the power transmission and transformation equipment; Correct the labeled infrared image according to the environmental detection parameters to generate an actual infrared image; Analyze the actual infrared image to generate equipment thermal image features, equipment temperature features, and equipment detection components; Determine the thermal imaging fault relationship and temperature defect relationship based on the device detection component and the preset device fault characteristic relationship; Obtain the device operating parameters of the device detection component; Determine the device detection report based on the device thermal imaging characteristics, thermal imaging fault relationship, device temperature characteristics, temperature defect relationship, and device operating parameters.

[0007] By adopting the above technical solution, after obtaining the labeled infrared image, the environmental detection parameters of the power transmission and transformation equipment are detected, and then the labeled infrared image is corrected according to the environmental detection parameters to obtain the actual infrared image, minimizing the influence of the external environment on the infrared image acquisition. The power transmission and transformation equipment can be monitored in more optional environments and at more times, thereby improving the timeliness of the power transmission and transformation equipment detection.

[0008] Optionally, the steps of correcting the labeled infrared image according to the environmental detection parameters to generate the actual infrared image include: Judge whether the environmental detection parameters meet the requirements of the preset ideal environmental parameters; If it meets the requirements, define the labeled infrared image as the actual infrared image; If it does not meet the requirements, analyze the environmental detection parameters and the ideal environmental parameters to determine the influencing environmental parameters; Correct the labeled infrared image according to the influencing environmental parameters to generate the actual infrared image.

[0009] By adopting the above technical solution, when the environmental detection parameters meet the requirements of the ideal environmental parameters, the labeled infrared image is directly defined as the actual infrared image. When they do not meet the requirements, the influencing environmental parameters in the environmental detection parameters are determined, and then the labeled infrared image is corrected according to the influencing environmental parameters, rather than correcting it with all environmental parameters, thereby improving the efficiency and accuracy of image correction.

[0010] Optionally, the steps of correcting the labeled infrared image according to the influencing environmental parameters to generate the actual infrared image include: Analyze the influencing environmental parameters to generate temperature compensation terms; the temperature compensation terms include environmental temperature compensation term, humidity temperature compensation term, wind speed temperature compensation term, and radiation temperature compensation term; Analyze the labeled infrared image to generate the initial detection temperature; Analyze the temperature compensation terms and the initial detection temperature to generate the true device temperature; Determine the true pixel value according to the true device temperature and the preset infrared temperature mapping relationship; Correct the labeled infrared image according to the true pixel value to generate the actual infrared image.

[0011] By adopting the above technical solution, after analyzing the environmental parameters that affect it, the temperature compensation term is determined. Then, after analyzing the temperature compensation term and the initial detection temperature, the true temperature of the device is determined. Next, according to the true temperature of the device, the true pixel value is found in the infrared temperature mapping relationship. Thus, the labeled infrared image is corrected with the true pixel value to obtain the actual infrared image, thereby improving the accuracy of the actual infrared image.

[0012] Optionally, the steps of analyzing the environmental parameters that affect it to generate the temperature compensation term include: Determine the temperature influence coefficient according to the actual target detection parameters and the preset target coefficient mapping relationship; Analyze the environmental parameters that affect it and the temperature influence coefficient to generate the temperature compensation term.

[0013] By adopting the above technical solution, the temperature influence coefficient is found in the target coefficient mapping relationship according to the actual target detection parameters. Thus, after calculating the environmental parameters that affect it and the temperature influence coefficient, the temperature compensation term is obtained, thereby improving the efficiency of determining the temperature compensation term.

[0014] Optionally, the steps of analyzing the environmental parameters that affect it and the temperature influence coefficient to generate the temperature compensation term include: Judge whether the environmental parameters that affect it include the preset wind speed environmental parameter or the preset radiation environmental parameter; If not, calculate the environmental parameters that affect it and the temperature influence coefficient according to the preset multiple regression model to generate the temperature compensation term; If so, analyze the environmental parameters that affect it to generate the general environmental parameter and the specific environmental parameter; Calculate the general environmental parameter and the temperature influence coefficient according to the preset multiple regression model to generate the general compensation term; Analyze the specific environmental parameter to generate the specific compensation term; Analyze the general compensation term and the specific compensation term to generate the temperature compensation term.

[0015] By adopting the above technical solution, when it is determined that the environmental parameters that affect it do not include the wind speed environmental parameter or the radiation environmental parameter, the temperature compensation term is directly calculated according to the multiple regression model for the environmental parameters that affect it and the temperature influence coefficient. When it includes the wind speed environmental parameter or the radiation environmental parameter, the environmental parameters that affect it are divided into the general environmental parameter and the specific environmental parameter. The general compensation term is calculated according to the multiple regression model for the general environmental parameter and the temperature influence coefficient. Then, after analyzing the specific environmental parameter, the specific compensation term is obtained. After correlating the general compensation term and the specific compensation term, the temperature compensation term is obtained, thereby improving the efficiency and accuracy of determining the temperature compensation term.

[0016] Optionally, the steps of analyzing the specific environmental parameter to generate the specific compensation term include: Determine whether specific environmental parameters meet the requirements of a preset specific environmental threshold; If they meet the requirements, calculate the specific environmental parameters and the temperature influence coefficient according to a multiple regression model to generate a specific compensation term; If they do not meet the requirements, calculate the specific environmental parameters and the temperature influence coefficient according to a preset hybrid model to generate a specific compensation term.

[0017] By adopting the above technical solution, when it is determined that the specific environmental parameters meet the requirements of the specific environmental threshold, it indicates that the influence of the specific environmental parameters on the temperature is still linear. Therefore, directly calculate the specific environmental parameters and the temperature influence coefficient according to the multiple regression model to obtain the specific compensation term. When they do not meet the requirements, it indicates that the influence of the specific environmental parameters on the temperature is in a non-linear stage. Therefore, calculate the specific environmental parameters and the temperature influence coefficient according to the hybrid model to determine the specific compensation term, thereby improving the accuracy of determining the specific compensation term.

[0018] Optionally, the steps of determining the device detection report according to the device thermal image characteristics, the thermal image fault relationship, the device temperature characteristics, the temperature defect relationship, and the device operating parameters include: Determine whether the device thermal image characteristics meet the requirements of the thermal image fault relationship; If they meet the requirements, determine the device fault characteristics according to the device thermal image characteristics and the thermal image fault relationship; If they do not meet the requirements, analyze the device operating parameters to determine the device fault characteristics; Determine the device defect characteristics according to the device temperature characteristics and the temperature defect relationship; Associate the device fault characteristics and the device defect characteristics to generate a device detection report.

[0019] By adopting the above technical solution, when it is determined that the device thermal image characteristics meet the requirements of the thermal image fault relationship, search for the device fault characteristics according to the device thermal image characteristics in the thermal image fault relationship. When they do not meet the requirements, determine the device fault characteristics according to the device operating parameters, and then search for the device defect characteristics according to the device temperature characteristics in the temperature defect relationship, so as to obtain the device detection report after associating the device fault characteristics and the device defect characteristics, thereby improving the accuracy of determining the device detection report.

[0020] Optionally, the steps of analyzing the device operating parameters to determine the device fault characteristics include: Determine whether the device operating parameters meet the requirements of preset normal operating parameters; If they meet the requirements, define the preset fault-free characteristics as the device fault characteristics; If they do not meet the requirements, determine the device fault characteristics according to the device operating parameters and the preset operating fault characteristic relationship.

[0021] By adopting the above technical solution, the operating parameters of the device are compared with the normal operating parameters. If the operating parameters of the device do not meet the requirements of the normal operating parameters, it indicates that the device has a fault. Therefore, the device fault characteristics are found according to the operating parameters of the device in the working fault characteristic relationship, so as to make up for the inability to determine the device fault through the infrared image with the operating parameters of the device, thereby improving the efficiency and accuracy of determining the device fault characteristics.

[0022] In a second aspect, the present application provides a power transmission and transformation equipment detection system based on deep learning, adopting the following technical solution: A power transmission and transformation equipment detection system based on deep learning, comprising: An acquisition module, configured to acquire an initial infrared image, a device detection image, and environmental detection parameters; A memory, configured to store the program of the power transmission and transformation equipment detection method based on deep learning as described in any one of the above; A processor, the program in the memory can be loaded and executed by the processor and implement the power transmission and transformation equipment detection method based on deep learning as described in any one of the above.

[0023] By adopting the above technical solution, the processor is controlled to load and execute the program of the power transmission and transformation equipment detection method stored in the memory, so that the acquisition module acquires a series of data related to the power transmission and transformation equipment detection based on deep learning. Thus, after obtaining the labeled infrared image, the environmental detection parameters of the power transmission and transformation equipment are detected, and the labeled infrared image is corrected according to the environmental detection parameters to obtain the actual infrared image, minimizing the influence of the external environment on the infrared image acquisition, and the power transmission and transformation equipment can be monitored in more optional environments and times, thereby improving the timeliness of the power transmission and transformation equipment detection.

[0024] In a third aspect, the present application provides an intelligent terminal, adopting the following technical solution: An intelligent terminal, comprising a memory and a processor, and a computer program capable of being loaded and executed by the processor as described in any one of the above power transmission and transformation equipment detection methods based on deep learning is stored on the memory.

[0025] By adopting the above technical solution, by operating the intelligent terminal, the processor is made to load and execute the computer program of the power transmission and transformation equipment detection method stored in the memory, so that after obtaining the labeled infrared image, the environmental detection parameters of the power transmission and transformation equipment are detected, and the labeled infrared image is corrected according to the environmental detection parameters to obtain the actual infrared image, minimizing the influence of the external environment on the infrared image acquisition, and the power transmission and transformation equipment can be monitored in more optional environments and times, thereby improving the timeliness of the power transmission and transformation equipment detection.

[0026] In summary, the present application includes at least one of the following beneficial technical effects: 1. After obtaining the labeled infrared image, the environmental detection parameters of the power transmission and transformation equipment are detected, and then the labeled infrared image is corrected according to the environmental detection parameters to obtain the actual infrared image, minimizing the influence of the external environment on the infrared image acquisition. The power transmission and transformation equipment can be monitored in more optional environments and at more times, thereby improving the timeliness of the power transmission and transformation equipment detection. 2. When the environmental detection parameters meet the requirements of the ideal environmental parameters, the labeled infrared image is directly defined as the actual infrared image. Otherwise, the environmental parameters affecting the environment are determined, and then the labeled infrared image is corrected according to the environmental parameters affecting the environment, rather than correcting it with all environmental parameters, thereby improving the efficiency and accuracy of image correction. 3. When it is determined that the specific environmental parameters meet the requirements of the specific environmental threshold, it indicates that the influence of the specific environmental parameters on the temperature is still linear. Therefore, the specific compensation term is directly calculated using the multiple regression model for the specific environmental parameters and the temperature influence coefficient. Otherwise, it indicates that the influence of the specific environmental parameters on the temperature is in a non-linear stage. Therefore, the specific compensation term is determined by calculating the specific environmental parameters and the temperature influence coefficient according to the hybrid model, thereby improving the accuracy of determining the specific compensation term. Description of the Drawings

[0027] Figure 1 is a flowchart of the method for detecting power transmission and transformation equipment based on deep learning in an embodiment of the present application.

[0028] Figure 2 is a flowchart of the steps for correcting the labeled infrared image according to the environmental detection parameters to generate the actual infrared image in an embodiment of the present application.

[0029] Figure 3 is a flowchart of the steps for correcting the labeled infrared image according to the environmental parameters affecting the environment to generate the actual infrared image in an embodiment of the present application.

[0030] Figure 4 is a flowchart of the steps for analyzing the environmental parameters affecting the environment to generate the temperature compensation term in an embodiment of the present application.

[0031] Figure 5 is a flowchart of the steps for analyzing the environmental parameters affecting the environment and the temperature influence coefficient to generate the temperature compensation term in an embodiment of the present application.

[0032] Figure 6 is a flowchart of the steps for analyzing the specific environmental parameters to generate the specific compensation term in an embodiment of the present application.

[0033] Figure 7It is a flowchart of the steps for determining an equipment detection report according to the equipment thermal image characteristics, thermal image fault relationships, equipment temperature characteristics, temperature defect relationships, and equipment operating parameters in the embodiments of the present application.

[0034] Figure 8 It is a flowchart of the steps for analyzing the equipment operating parameters to determine the equipment fault characteristics in the embodiments of the present application. Detailed implementation manners

[0035] In order to make the objectives, technical solutions, and advantages of the present application clearer and more understandable, the following further describes the present application in detail with reference to the Figures 1 to 8 accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0036] The embodiments of the present application disclose a power transmission and transformation equipment detection method based on deep learning. Specifically, a front-end infrared intelligent thermal imager with an infrared thermal imaging module, an image acquisition module, a fault diagnosis module, and a processing terminal is disclosed. The initial infrared image of the power transmission and transformation equipment and the equipment detection image are respectively acquired through the infrared thermal imaging module and the image acquisition module. Thus, the processing terminal inputs the equipment detection image into the target detection model to analyze the image target detection parameters, and then adjusts the image target detection parameters according to the image deviation parameters between the infrared thermal imaging module and the image acquisition module to obtain the actual target detection parameters. Thus, the actual target detection parameters are marked on the initial infrared image to form a marked infrared image. After detecting the environmental detection parameters of the power transmission and transformation equipment, the processing terminal corrects the marked infrared image according to the environmental detection parameters to obtain the actual infrared image. Thus, an equipment detection report is obtained after analyzing the actual infrared image, and the infrared image detected by the infrared thermal imager will not be inaccurate due to environmental factors, so there are more optional times for detection, thereby improving the timeliness of power transmission and transformation equipment detection.

[0037] Referring to Figure 1 , the embodiments of the present application disclose a power transmission and transformation equipment detection method based on deep learning, including the following steps: Step S100: Obtain the initial infrared image and the equipment detection image of the power transmission and transformation equipment.

[0038] Among them, the initial infrared image refers to the infrared image of the power transmission and transformation equipment directly acquired by the infrared thermal imaging module. The infrared thermal imaging module uses infrared thermal imaging technology to detect the infrared radiation of the power transmission and transformation equipment, and through means such as photoelectric conversion and signal processing, converts the temperature distribution map of the power transmission and transformation equipment into a video image.

[0039] The device detection image refers to the image of the power transmission and transformation equipment collected by the image acquisition module. The image acquisition module can use a camera. In the embodiments of the present application, the image acquisition module and the infrared imaging module are installed side by side to reduce the error of the target position in the initial infrared image and the device detection image.

[0040] Step S101: Control a preset target detection model to analyze the device detection image to generate image target detection parameters.

[0041] Among them, the target detection model refers to a deep learning model used to locate the position and type of detection targets in an image. In the embodiments of the present application, taking the Faster RCNN deep target detection neural network as an example, first establish an image sample library of power transmission and transformation equipment, use the training set of the sample library to train the established Faster RCNN deep target detection neural network, verify the overfitting degree of the model through the validation set, and then use the network model established by training to perform multi-target recognition and positioning on the images in the test set to generate the recognition results of the target position and type; The Faster RCNN algorithm consists of two major modules: the RPN candidate box extraction module and the Faster RCNN detection module. Among them, RPN is a fully convolutional neural network used to extract candidate boxes; Faster RCNN detects and identifies the targets in the proposal based on the proposal extracted by RPN.

[0042] The image target detection parameters refer to the position and type of detection targets in the image, which are obtained by inputting the device detection image into the target detection model by the processing terminal for recognition. The reason for performing target recognition and detection on the device detection image instead of the initial infrared image is that the resolution of the device detection image is higher, which is easier for the model to perform target recognition and the recognition accuracy is higher.

[0043] Step S102: Correct the image target detection parameters according to the preset image deviation parameters to generate actual target detection parameters.

[0044] Among them, the image deviation parameter refers to the position deviation of the target in the images of both the image acquisition module and the infrared thermal imaging module. The operator compares the position deviation of the target in the images collected by the image acquisition module and the infrared thermal imaging module and stores the position deviation in the processing terminal.

[0045] The actual target detection parameters refer to the position and type of the target in the infrared image, which are obtained by the processing terminal adjusting the target position in the image target detection parameters according to the position deviation corresponding to the image deviation parameter.

[0046] Step S103: Label the initial infrared image according to the actual target detection parameters to generate a labeled infrared image.

[0047] Among them, the labeled infrared image refers to the infrared image labeled with the target bounding box and type. The processing terminal reads the target position and type in the actual target detection parameters, and then uses the drawing function provided by the image processing library to draw the bounding box on the initial infrared image and add the type annotation to obtain the labeled infrared image.

[0048] Step S104: Obtain the environmental detection parameters of the power transmission and transformation equipment.

[0049] Among them, the environmental detection parameters refer to the parameters of the environment where the power transmission and transformation equipment is located, including parameters such as environmental temperature, humidity, wind speed, and solar radiation. They are detected by a distributed temperature and humidity sensor network, a non-contact solar radiation intensity sensor, and an ultrasonic anemometer. The environmental temperature directly affects the heat exchange between the equipment surface and the environment, the wind speed affects the heat dissipation efficiency of the equipment, the humidity affects the absorption in the infrared band, and the solar radiation causes additional heating on the equipment surface, all of which may cause the infrared temperature measurement value to deviate from the true temperature.

[0050] Step S105: Correct the labeled infrared image according to the environmental detection parameters to generate the actual infrared image.

[0051] Among them, the actual infrared image refers to the true infrared image of the power transmission and transformation equipment without being affected by the environment. It is generated by the processing terminal after correcting the labeled infrared image according to the environmental detection parameters. The specific method refers to Figure 2 the steps.

[0052] Step S106: Analyze the actual infrared image to generate the device thermal image feature, the device temperature feature, and the device detection component.

[0053] Among them, the device thermal image feature refers to the thermal image feature of the target in the actual infrared image. For example, the fault thermal image feature of the current transformer includes overall heating centered on the body, and the fault thermal image feature of the voltage transformer is that the overall temperature rise is high, and the temperature in the upper middle part is high, etc. It is obtained by the processing terminal using the trained convolutional neural network to extract the actual infrared image.

[0054] The device temperature feature refers to the temperature of the device. The processing terminal extracts the pixel values within the bounding box in the actual infrared image, then looks up the actual temperature in the corresponding relationship between the pixel and the temperature according to the pixel values, and then calculates the temperature difference and relative temperature difference between the actual temperatures to obtain the device temperature feature.

[0055] The device detection component refers to the specific component to be detected, which is obtained by the processing terminal identifying the annotation in the actual infrared image.

[0056] Step S107: Determine the thermal image fault relationship and the temperature defect relationship according to the device detection component and the preset device fault feature relationship.

[0057] Among them, the relationship between equipment fault characteristics refers to the corresponding relationship between different equipment and the relationship between thermal imaging faults and temperature defects, which is summarized by the operator according to the defect diagnosis basis in DL / T664 "Technical Specification for Infrared Diagnosis of Energized Equipment". For example, the thermal imaging characteristics of metal wires are the thermal images centered on the wires, with obvious hot spots. Then the fault characteristics are loose strands, broken strands, aging or insufficient cross-sectional area. When the relative temperature difference reaches 35%, but the hot spot temperature does not reach the threshold, it is a general defect, etc.

[0058] The relationship between thermal imaging faults refers to the fault characteristics corresponding to different thermal images of the equipment, and the relationship between temperature defects refers to the nature of the defects corresponding to different temperature characteristics of the equipment, which are found by the processing terminal in the relationship between equipment fault characteristics according to the component type corresponding to the equipment detection component.

[0059] Step S108: Obtain the equipment operating parameters of the equipment detection component.

[0060] Among them, the equipment operating parameters refer to the parameters when the equipment detection component works, such as volt-ampere characteristics, resistance, etc., which are sent from the SCADA system to the processing terminal.

[0061] Step S109: Determine the equipment detection report according to the equipment thermal imaging characteristics, the relationship between thermal imaging faults, the equipment temperature characteristics, the relationship between temperature defects, and the equipment operating parameters.

[0062] Among them, the equipment detection report refers to the report of equipment detection faults, including fault characteristics and defect levels. The fault characteristics include poor contact, loose strands, broken strands, and aging, etc. The defect levels include general defects, serious defects, and emergency defects, which are determined by the processing terminal after analyzing the equipment thermal imaging characteristics, the relationship between thermal imaging faults, the equipment temperature characteristics, the relationship between temperature defects, and the equipment operating parameters. The specific method refers to Figure 7 the steps.

[0063] Refer to Figure 2 , the steps of correcting the labeled infrared image according to the environmental detection parameters to generate the actual infrared image include: Step S200: Determine whether the environmental detection parameters meet the requirements of the preset ideal environmental parameters.

[0064] Among them, the ideal environmental parameters refer to the environmental parameters with the directly detected infrared image as the actual image. For example, the environmental temperature is 0 degrees Celsius, the relative humidity is 85%, the wind speed is 5 m / s, and the solar radiation intensity is 0 W / m 2 , the requirements for the ideal environmental parameters refer to not exceeding the error range of the ideal environmental parameters, and the error range is determined by the operator according to the actual situation.

[0065] The processing terminal determines whether the environmental detection parameters do not exceed the error range of the ideal environmental parameters, so as to determine whether the labeled infrared image collected by the infrared thermal imaging module can be directly used.

[0066] Step S201: If it meets the requirements, define the labeled infrared image as the actual infrared image.

[0067] Among them, if the processing terminal determines that the environmental detection parameters do not exceed the error range of the ideal environmental parameters, it indicates that the environment where the power transmission and transformation equipment is located has little influence on the temperature of the power transmission and transformation equipment, and the infrared image collected by the infrared thermal imaging module is the true temperature image of the power transmission and transformation equipment. Therefore, the labeled infrared image is defined as the actual infrared image.

[0068] Step S202: If it does not meet the requirements, analyze the environmental detection parameters and the ideal environmental parameters to determine the influencing environmental parameters.

[0069] Among them, if the processing terminal determines that the environmental detection parameters exceed the error range of the ideal environmental parameters, it indicates that the environment where the power transmission and transformation equipment is located has an impact on the true temperature of the power transmission and transformation equipment, which may cause the temperature of the power transmission and transformation equipment to be higher than the actual working temperature or may cause the temperature of the power transmission and transformation equipment to be lower than the actual working temperature. Therefore, after analyzing the environmental detection parameters and the ideal environmental parameters, determine the influencing environmental parameters to provide data support for subsequent correction of the labeled infrared image.

[0070] The influencing environmental parameters refer to the environmental parameters that will affect the temperature of the power transmission and transformation equipment. The processing terminal compares the environmental detection parameters with the ideal environmental parameters and defines the environmental detection parameters that exceed the error range of the ideal environmental parameters as the influencing environmental parameters.

[0071] Step S2021: Correct the labeled infrared image according to the influencing environmental parameters to generate the actual infrared image.

[0072] Among them, the actual infrared image in this step is the same as the actual infrared image in Step S105. The processing terminal determines the influencing temperature after analyzing the influencing environmental parameters, and then corrects the pixels of the labeled infrared image with the influencing temperature to obtain the actual infrared image. The specific method refers to Figure 3 the steps.

[0073] Refer to Figure 3 , the steps of correcting the labeled infrared image according to the influencing environmental parameters to generate the actual infrared image include: Step S300: Analyze the influencing environmental parameters to generate temperature compensation terms; the temperature compensation terms include environmental temperature compensation terms, humidity temperature compensation terms, wind speed temperature compensation terms, and radiation temperature compensation terms.

[0074] Among them, the temperature compensation item refers to the temperature of the environment's impact on power transmission and transformation equipment, including the environmental temperature compensation item, humidity-temperature compensation item, wind speed-temperature compensation item, and radiation temperature compensation item, which are determined by the processing terminal through analysis and calculation of the environmental parameter impacts. For the specific method, refer to Figure 4 the steps.

[0075] Step S301: Analyze the labeled infrared image to generate an initial detected temperature.

[0076] Among them, the initial detected temperature refers to the temperature of the power transmission and transformation equipment detected by the infrared thermal imaging module, and the processing terminal obtains the initial detected temperature by looking up the pixels within the bounding box in the labeled infrared image in the corresponding relationship between pixel sum and temperature.

[0077] Step S302: Analyze the temperature compensation item and the initial detected temperature to generate the true equipment temperature.

[0078] Among them, the true equipment temperature refers to the true temperature of the power transmission and transformation equipment after removing the environmental impact, which is determined by the processing terminal calculating the sum of the temperature compensation item and the initial detected temperature.

[0079] Step S303: Determine the true pixel value according to the true equipment temperature and the preset infrared temperature mapping relationship.

[0080] Among them, the infrared temperature mapping relationship refers to the corresponding relationship between pixel value and temperature, and the operator forms a mapping table by corresponding different temperatures to different pixel values one by one.

[0081] The true pixel value refers to the pixel value of the true temperature of the power transmission and transformation equipment, which is obtained by the processing terminal looking up in the infrared temperature mapping relationship according to the true equipment temperature.

[0082] Step S304: Correct the labeled infrared image according to the true pixel value to generate the actual infrared image.

[0083] Among them, the actual infrared image in this step is the same as the actual infrared image in Step S2021. The processing terminal uses the image processing library to correct the pixel value of the standard infrared image to the true pixel value, thereby obtaining the actual infrared image, which reflects the true temperature of the power transmission and transformation equipment.

[0084] Refer to Figure 4 , the steps of analyzing the environmental parameter impacts to generate the temperature compensation item include: Step S400: Determine the temperature impact coefficient according to the actual target detection parameters and the preset target coefficient mapping relationship.

[0085] Among them, the target coefficient mapping relationship refers to the temperature influence coefficients corresponding to different types of targets, including the ambient temperature influence coefficient, humidity influence coefficient, wind speed influence coefficient, and solar radiation influence coefficient. These coefficients are obtained through calibration tests on different power transmission and transformation equipment by operators, and a mapping table is formed by corresponding the target types with the temperature influence coefficients one by one.

[0086] The temperature influence coefficient refers to the coefficient by which the temperature of power transmission and transformation equipment is affected by the environment, including the ambient temperature influence coefficient, humidity influence coefficient, wind speed influence coefficient, and solar radiation influence coefficient. It is obtained by the processing terminal searching in the mapping table corresponding to the target coefficient mapping relationship according to the target type in the actual target detection parameters.

[0087] Step S401: Analyze the environmental parameters and temperature influence coefficients to generate a temperature compensation term.

[0088] Among them, the temperature compensation term in this step is the same as the temperature compensation term in step S300, and is obtained by the processing terminal calculating the environmental parameters and temperature influence coefficients. For the specific method, refer to Figure 5 the steps.

[0089] Refer to Figure 5 , the steps of analyzing the environmental parameters and temperature influence coefficients to generate a temperature compensation term include: Step S500: Determine whether the environmental parameters include a preset wind speed environmental parameter or a preset radiation environmental parameter.

[0090] Among them, the wind speed environmental parameter means that the environmental parameter is wind speed, and the radiation environmental parameter means that the environmental parameter is solar radiation, which are stored in the processing terminal by the operator.

[0091] The processing terminal determines whether the environmental parameters include a wind speed environmental parameter or a radiation environmental parameter, so as to determine whether the temperature compensation term can be directly calculated.

[0092] Step S501: If not, calculate the environmental parameters and temperature influence coefficients according to a preset multiple regression model to generate a temperature compensation term.

[0093] Among them, if the processing terminal determines that the environmental parameters do not include a wind speed environmental parameter or a radiation environmental parameter, it indicates that the influence of the environmental parameters on the temperature of the power transmission and transformation equipment is linear. Therefore, the environmental parameters and temperature influence coefficients are input into the multiple regression model to calculate the temperature compensation term.

[0094] The multiple regression model refers to the model used to calculate the compensated temperature, including the temperature compensation model, humidity compensation model, wind speed compensation model, and solar radiation compensation model. The formula of the temperature compensation model is the difference between the ambient temperature and the ideal ambient temperature multiplied by the ambient temperature influence coefficient. The formula of the humidity compensation model is the humidity influence coefficient multiplied by the relative ambient humidity. The formula of the wind speed compensation model is the negative exponential form of the ambient wind speed multiplied by the wind speed influence coefficient, reflecting that the heat dissipation efficiency slows down with the increase of wind speed. The formula of the solar radiation compensation model is the solar radiation coefficient multiplied by the solar radiation intensity.

[0095] The temperature compensation term in this step is the same as the temperature compensation term in step S401. The processing terminal calls the corresponding model in the multiple regression model according to the environmental parameters affecting it, so as to input the environmental parameters affecting it and the temperature influence coefficient into the model for calculation.

[0096] Step S502: If so, analyze the environmental parameters affecting it to generate ordinary environmental parameters and specific environmental parameters.

[0097] Among them, if the processing terminal determines that the environmental parameters affecting it include wind speed environmental parameters or radiation environmental parameters, it indicates that the environment has linear and non-linear effects on the temperature of the power transmission and transformation equipment. Therefore, after analyzing the environmental parameters affecting it, ordinary environmental parameters and specific environmental parameters are determined to provide data support for subsequent calculation of the temperature compensation term.

[0098] Ordinary environmental parameters refer to the parameters in the environmental parameters affecting it that do not cause non-linear effects, that is, the environmental temperature parameter and the humidity parameter, which are obtained by the processing terminal identifying the environmental temperature parameter or the relative humidity parameter in the environmental parameters affecting it.

[0099] Specific environmental parameters refer to the parameters in the environmental parameters affecting it that may cause non-linear effects, that is, the wind speed parameter and the solar radiation intensity, which are obtained by the processing terminal identifying the wind speed parameter or the solar radiation intensity in the environmental parameters affecting it.

[0100] Step S503: Calculate the ordinary environmental parameters and the temperature influence coefficient according to the preset multiple regression model to generate an ordinary compensation term.

[0101] Among them, the multiple regression model in this step is the same as the multiple regression model in step S501, which will not be elaborated here.

[0102] The ordinary compensation term refers to the temperature compensation caused by the ordinary environmental parameters to the power transmission and transformation equipment. The processing terminal selects the corresponding model from the multiple regression model according to the ordinary environmental parameters, and then inputs the ordinary environmental parameters and the temperature influence coefficient into the model to calculate the ordinary compensation term.

[0103] Step S504: Analyze the specific environmental parameters to generate a specific compensation term.

[0104] Among them, the specific compensation item refers to the temperature compensation caused by specific environmental parameters to the power transmission and transformation equipment, which is determined by the processing terminal after analyzing the specific environmental parameters. For the specific method, refer to Figure 6 the steps of

[0105] Step S505: Analyze the general compensation item and the specific compensation item to generate a temperature compensation item.

[0106] Among them, the temperature compensation item in this step is the same as the temperature compensation item in Step S401, and is obtained by the processing terminal calculating the sum of the general compensation item and the specific compensation item.

[0107] Refer to Figure 6 , the steps of analyzing the specific environmental parameters to generate a specific compensation item include: Step S600: Determine whether the specific environmental parameters meet the requirements of the preset specific environmental threshold.

[0108] Among them, the specific environmental threshold refers to the threshold at which the environmental impact on temperature enters the non-linear stage. For example, the wind speed is 10 m / s and the solar radiation intensity is 1000 W / m 2 , and the requirement of the specific environmental threshold means not exceeding the specific environmental threshold.

[0109] The processing terminal determines whether the specific environmental parameters do not exceed the specific environmental threshold, so as to determine whether the impact of the specific environment on the power transmission and transformation equipment reaches the non-linear stage.

[0110] Step S601: If it meets the requirements, calculate the specific environmental parameters and the temperature influence coefficient according to the multiple regression model to generate a specific compensation item.

[0111] Among them, if the processing terminal determines that the specific environmental parameters do not exceed the specific environmental threshold, it indicates that the impact of the specific environment on the power transmission and transformation equipment has not reached the non-linear stage. Therefore, the specific compensation item is still obtained by calculating the specific environmental parameters and the temperature influence coefficient according to the model in the multiple regression model.

[0112] Step S602: If it does not meet the requirements, calculate the specific environmental parameters and the temperature influence coefficient according to the preset hybrid model to generate a specific compensation item.

[0113] Among them, if the processing terminal determines that the specific environmental parameters exceed the specific environmental threshold, it indicates that the impact of the specific environment on the power transmission and transformation equipment reaches the non-linear stage. Therefore, the specific compensation item is obtained by calculating the specific environmental parameters and the temperature influence coefficient according to the hybrid model.

[0114] The hybrid model refers to the model obtained by adding non-linear factors to the original wind speed compensation model and solar radiation compensation model. The formula of the hybrid model is the original model formula plus the square of the original model formula. For example, A + A 2 .

[0115] Referring to Figure 7 , the steps to determine the equipment inspection report according to the equipment thermal image characteristics, thermal image fault relationship, equipment temperature characteristics, temperature defect relationship, and equipment operating parameters include: Step S700: Determine whether the equipment thermal image characteristics meet the requirements of the thermal image fault relationship.

[0116] Among them, the requirements of the thermal image fault relationship refer to those existing in the thermal image corresponding to the thermal image fault relationship. The processing terminal determines whether the equipment thermal image characteristics exist in the thermal image corresponding to the thermal image fault relationship, so as to determine whether the equipment has a fault.

[0117] Step S701: If it meets the requirements, determine the equipment fault characteristics according to the equipment thermal image characteristics and the thermal image fault relationship.

[0118] Among them, if the processing terminal determines that the equipment thermal image characteristics exist in the thermal image corresponding to the thermal image fault relationship, it indicates that the equipment has a fault. Therefore, determine the equipment fault characteristics according to the equipment thermal image characteristics and the thermal image fault relationship.

[0119] The equipment fault characteristics refer to the specific faults that occur in the equipment, which are obtained by the processing terminal searching in the mapping table corresponding to the thermal image fault relationship according to the equipment thermal image characteristics.

[0120] Step S702: If it does not meet the requirements, analyze the equipment operating parameters to determine the equipment fault characteristics.

[0121] Among them, if the processing terminal determines that the equipment thermal image characteristics do not exist in the thermal image corresponding to the thermal image fault relationship, it indicates that the equipment may or may not have a fault. Therefore, the equipment status cannot be determined from the thermal image characteristics. Thus, analyze the equipment operating parameters of the equipment detection component to provide data support for subsequent determination of the equipment fault characteristics.

[0122] The equipment fault characteristics in this step are defined in the same way as the equipment fault characteristics in Step S701, and are determined by the processing terminal after analyzing the equipment operating parameters. The specific method refers to Figure 8 the steps of

[0123] Step S703: Determine the equipment defect characteristics according to the equipment temperature characteristics and the temperature defect relationship.

[0124] Among them, the device defect feature refers to the nature of the device defect, including general defects, serious defects, and urgent defects, which is obtained by the processing terminal looking up in the mapping table corresponding to the temperature defect relationship according to the device temperature feature.

[0125] Step S704: Associate the device fault feature and the device defect feature to generate a device detection report.

[0126] Among them, the device detection report in this step is the same as the device detection report in step S109, which is formed by the processing terminal storing the fault feature corresponding to the device fault feature and the defect nature corresponding to the device defect feature in the same database.

[0127] Refer to Figure 8 , the steps of analyzing the device working parameters to determine the device fault feature include: Step S800: Judge whether the device working parameters meet the requirements of the preset normal working parameters.

[0128] Among them, the normal working parameters refer to the parameters when the device is working normally, such as the volt-ampere characteristic, etc., which are obtained by the operator testing the normal device. The requirements for the normal working parameters refer to within the parameter range of the normal working parameters.

[0129] The processing terminal judges whether the device working parameters are within the parameter range of the normal working parameters, so as to determine whether the device detection component is in a normal state.

[0130] Step S801: If it meets the requirements, define the preset no-fault feature as the device fault feature.

[0131] Among them, if the processing terminal determines that the device working parameters are within the parameter range of the normal working parameters, it indicates that the device detection component is in a normal state. Therefore, the no-fault feature is defined as the device fault feature.

[0132] The no-fault feature means that the device does not have any fault features, which are stored in the processing terminal by the operator.

[0133] Step S802: If it does not meet the requirements, determine the device fault feature according to the device working parameters and the preset working fault feature relationship.

[0134] Among them, if the processing terminal determines that the device working parameters are not within the parameter range of the normal working parameters, it indicates that the device detection component is in a fault state. Therefore, the device fault feature is obtained by looking up in the mapping table corresponding to the working fault feature relationship according to the device working parameters.

[0135] The working fault feature relationship refers to the corresponding relationship between working parameters and fault features. When a fault occurs in the device detection component, the operator detects the working parameters at the corresponding moment and forms a mapping table by corresponding the fault features with the working parameters one by one.

[0136] Based on the same inventive concept, an embodiment of the present application provides a power transmission and transformation equipment detection system based on deep learning, including: An acquisition module, configured to acquire an initial infrared image, a device detection image, environmental detection parameters, and device working parameters; A memory, configured to store a program of a power transmission and transformation equipment detection method based on deep learning; A processor, the program in the memory can be loaded and executed by the processor and implement a power transmission and transformation equipment detection method based on deep learning.

[0137] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the above division of each functional module is used as an example. In actual applications, the above functions can be allocated to different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. The specific working processes of the above-described system, device, and unit can refer to the corresponding processes in the foregoing method embodiments and will not be described herein again.

[0138] An embodiment of the present application provides a computer-readable storage medium, storing a computer program that can be loaded and executed by a processor to implement a power transmission and transformation equipment detection method based on deep learning.

[0139] Computer storage media include, for example: various media such as USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical discs that can store program codes.

[0140] Based on the same inventive concept, an embodiment of the present application provides an intelligent terminal, including a memory and a processor, and a computer program that can be loaded and executed by the processor to implement a power transmission and transformation equipment detection method based on deep learning is stored on the memory.

[0141] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the above division of each functional module is used as an example. In actual applications, the above functions can be allocated to different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. The specific working processes of the above-described system, device, and unit can refer to the corresponding processes in the foregoing method embodiments and will not be described herein again.

[0142] The above are all preferred embodiments of the present application, and do not limit the protection scope of the present application accordingly. Any feature disclosed in this specification (including the abstract and drawings), unless specifically described, can be replaced by other equivalent or similar-purpose alternative features. That is, unless specifically described, each feature is only an example of a series of equivalent or similar features.

Claims

1. A power transmission and transformation equipment detection method based on deep learning, characterized in that: include: Obtaining initial infrared images and equipment detection images of power transmission and transformation equipment; Control the preset target detection model to analyze the device detection image to generate image target detection parameters; Correcting the image target detection parameters according to the preset image deviation parameters to generate actual target detection parameters; Annotating the initial infrared image according to actual target detection parameters to generate an annotated infrared image; Obtain environmental detection parameters of power transmission and transformation equipment; Correcting the annotated infrared image according to the environmental detection parameters to generate an actual infrared image; Analyze the actual infrared image to generate equipment thermal image characteristics, equipment temperature characteristics and equipment detection components; Determine the thermal image fault relationship and temperature defect relationship based on the equipment detection components and the preset equipment fault characteristic relationship; Obtain equipment working parameters of equipment detection components; The equipment inspection report is determined based on the equipment thermal imaging characteristics, thermal imaging fault relationships, equipment temperature characteristics, temperature defect relationships and equipment operating parameters.

2. The power transmission and transformation equipment detection method based on deep learning according to claim 1 is characterized in that: The steps of correcting the annotated infrared image according to the environmental detection parameters to generate the actual infrared image include: Determine whether the environmental detection parameters meet the requirements of the preset ideal environmental parameters; If it is in compliance, the annotated infrared image is defined as the actual infrared image; If not, analyze the environmental detection parameters and ideal environmental parameters to determine the influencing environmental parameters; The annotated infrared image is corrected according to the influencing environmental parameters to generate the actual infrared image.

3. The power transmission and transformation equipment detection method based on deep learning according to claim 2 is characterized in that: The steps of correcting the annotated infrared image according to the influencing environment parameters to generate the actual infrared image include: Analyze the parameters affecting the environment to generate temperature compensation items; the temperature compensation items include ambient temperature compensation items, humidity temperature compensation items, wind speed temperature compensation items and radiation temperature compensation items; Analyzing the annotated infrared image to generate an initial detection temperature; Analyze the temperature compensation term and the initial detection temperature to generate the real temperature of the device; Determine the real pixel value based on the real temperature of the device and the preset infrared temperature mapping relationship; The annotated infrared image is corrected according to the real pixel value to generate the actual infrared image.

4. The power transmission and transformation equipment detection method based on deep learning according to claim 3 is characterized in that: The steps of analyzing the influencing environmental parameters to generate temperature compensation terms include: Determine the temperature influence coefficient according to the mapping relationship between the actual target detection parameters and the preset target coefficient; The influencing environmental parameters and temperature influence coefficients are analyzed to generate temperature compensation items.

5. The power transmission and transformation equipment detection method based on deep learning according to claim 4 is characterized in that: The steps of analyzing the influencing environmental parameters and temperature influence coefficients to generate temperature compensation items include: Determine whether the influencing environmental parameters include preset wind speed environmental parameters or preset radiation environmental parameters; If not, the influencing environmental parameters and temperature influence coefficients are calculated according to a preset multivariate regression model to generate a temperature compensation term; If yes, the influencing environmental parameters are analyzed to generate general environmental parameters and specific environmental parameters; Calculate common environmental parameters and temperature influence coefficients according to a preset multiple regression model to generate common compensation terms; Analyze specific environmental parameters to generate specific compensation items; The general compensation terms and the specific compensation terms are analyzed to generate the temperature compensation terms.

6. The power transmission and transformation equipment detection method based on deep learning according to claim 5 is characterized in that: The steps for analyzing specific environmental parameters to generate specific compensation terms include: Determine whether specific environmental parameters meet the requirements of preset specific environmental thresholds; If it is met, the specific environmental parameters and temperature influence coefficients are calculated according to the multivariate regression model to generate specific compensation items; If not, the specific environmental parameters and temperature influence coefficients are calculated according to a preset hybrid model to generate specific compensation items.

7. The power transmission and transformation equipment detection method based on deep learning according to claim 1, characterized in that: The steps of determining the equipment inspection report based on the equipment thermal image characteristics, thermal image fault relationship, equipment temperature characteristics, temperature defect relationship and equipment operating parameters include: Determine whether the thermal imaging characteristics of the equipment meet the requirements of thermal imaging fault relationship; If it meets the requirements, the equipment failure characteristics are determined based on the equipment thermal image characteristics and thermal image failure relationship; If not, analyze the equipment operating parameters to determine the equipment failure characteristics; Determine equipment defect characteristics based on equipment temperature characteristics and temperature defect relationships; Correlate equipment failure characteristics and equipment defect characteristics to generate equipment inspection reports.

8. The power transmission and transformation equipment detection method based on deep learning according to claim 1, characterized in that: The steps for analyzing the equipment operating parameters to determine the equipment failure characteristics include: Determine whether the equipment operating parameters meet the preset normal operating parameters; If it meets the requirement, the preset no-fault feature is defined as the equipment fault feature; If not, the equipment fault characteristics are determined based on the equipment operating parameters and the preset working fault characteristic relationship.

9. A power transmission and transformation equipment detection system based on deep learning, characterized in that: include: An acquisition module is used to acquire an initial infrared image, a device detection image, and environmental detection parameters; A memory, used to store a program of a power transmission and transformation equipment detection method based on deep learning according to any one of claims 1 to 8; The program in the processor memory can be loaded and executed by the processor to implement the power transmission and transformation equipment detection method based on deep learning as described in any one of claims 1 to 8.

10. An intelligent terminal, characterized in that: It comprises a memory and a processor, wherein the memory stores a computer program that can be loaded by the processor and executes the deep learning-based power transmission and transformation equipment detection method as described in any one of claims 1 to 8.

Images