Power equipment temperature identification method, system, terminal and storage medium

Abstract

The application provides a power equipment temperature identification method, system, terminal and storage medium, and the method comprises the following steps: performing keyword detection on the equipment infrared image of the power equipment to be identified to obtain a keyword detection result; performing region cropping according to the keyword detection result to obtain a temperature value region, performing gray scale conversion on the temperature value region to obtain a temperature gray scale image; performing pixel value conversion on the temperature gray scale image, and performing binaryzation processing on the temperature gray scale image after the pixel value conversion; inputting the temperature gray scale image after the binaryzation processing into a temperature value identification model to perform temperature value identification, and obtaining the highest temperature value and the lowest temperature value of the power equipment to be identified. The temperature value identification model can automatically perform temperature value identification on the temperature gray scale image after the binaryzation processing to obtain the highest temperature value and the lowest temperature value, so that the reading and recording of the highest temperature value and the lowest temperature value are not required in an artificial manner, and the power equipment temperature identification efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power equipment detection, and in particular to a method, system, terminal and storage medium for identifying the temperature of power equipment. Background Art

[0002] Electricity is one of the important resources to maintain people's normal life. With the development of social economy, domestic electricity consumption is constantly increasing. Ensuring the safe and stable operation of the power system has become an increasingly important task. The power system may include multiple power equipment. The temperature identification of power equipment is a key process for fever detection and fault diagnosis. Therefore, the temperature identification method of power equipment is increasingly valued by people.

[0003] In the existing temperature recognition process of power equipment, the highest and lowest temperature values ​​in the infrared image of the equipment are generally recorded manually, resulting in a large waste of manpower and material resources and reducing the efficiency of power equipment temperature recognition. Summary of the Invention

[0004] The purpose of the embodiments of the present invention is to provide a method, system, terminal and storage medium for identifying the temperature of electric equipment, aiming to solve the problem of low efficiency of existing methods for identifying the temperature of electric equipment.

[0005] The embodiment of the present invention is implemented as follows: a method for identifying the temperature of an electric power device, the method comprising:

[0006] Acquire an infrared image of the power equipment to be identified, and perform keyword detection on the infrared image of the equipment to obtain a keyword detection result;

[0007] Performing region cropping on the infrared image of the device according to the keyword detection result to obtain a temperature value region, and performing grayscale transformation on the temperature value region to obtain a temperature grayscale map;

[0008] Performing pixel value conversion on the temperature grayscale image, and performing binarization processing on the temperature grayscale image after pixel value conversion;

[0009] The binarized temperature grayscale image is input into a pre-trained temperature value recognition model to perform temperature value recognition, thereby obtaining the maximum temperature value and the minimum temperature value of the power equipment to be recognized.

[0010] Furthermore, the keyword detection on the infrared image of the device includes:

[0011] Performing color bar detection on the infrared image of the device according to the pre-trained target detection model;

[0012] If the color mark bar is detected in the infrared image of the device, obtaining the location area of ​​the color mark bar and obtaining the keyword detection result;

[0013] If the color mark bar is not detected in the device infrared image, performing preset character detection on the device infrared image according to the target detection model, the preset characters including a first character and a second character;

[0014] If the preset character is detected in the infrared image of the device, the position area of ​​the first character and the position area of ​​the second character are acquired to obtain the keyword detection result.

[0015] Furthermore, the performing of region cropping on the device infrared image according to the keyword detection result to obtain a temperature value region includes:

[0016] If the keyword detection result includes the location area of ​​the color mark bar, determining the area above and below the location area of ​​the color mark bar by a first preset distance as the cropping area;

[0017] If the keyword detection result includes the position area of ​​the first character and the position area of ​​the second character, obtaining the position relationship between the first character and the second character;

[0018] If the first character and the second character are in a horizontal relationship, an area at a second preset distance below the position area of ​​the first character and the position area of ​​the second character is determined as the cropping area;

[0019] If the first character and the second character are in a vertical relationship, an area at a third preset distance to the right of the position area of ​​the first character and the position area of ​​the second character is determined as the cropping area;

[0020] In the infrared image of the device, the cropping area is cropped to obtain the temperature value area.

[0021] Furthermore, after performing keyword detection on the infrared image of the device and obtaining the keyword detection result, the method further includes:

[0022] If the color bar and the preset characters are not detected in the infrared image of the device, the device type and voltage level of the power device to be identified are matched with historical infrared data to obtain matching data;

[0023] Obtaining shooting year and month information of the infrared image of the device, and filtering the matching data according to the shooting year and month information to obtain first filtered data;

[0024] Acquire the device address and substation name of the power device to be identified, and filter the first filtering data according to the device address and the substation name to obtain second filtering data;

[0025] Obtaining an image capture time of the device infrared image, and respectively calculating a time difference between a historical capture time of each historical infrared image in the second filtered data and the image capture time;

[0026] The historical infrared image corresponding to the minimum time difference is determined as the target infrared image, and the highest temperature value and the lowest temperature value of the target infrared image are determined as the highest temperature value and the lowest temperature value of the power equipment to be identified.

[0027] Furthermore, the pixel value conversion of the temperature grayscale image includes:

[0028] Obtain the number of occurrences of each pixel value in the temperature grayscale image respectively, and calculate the average value between the two pixel values ​​corresponding to the maximum number of occurrences to obtain a pixel threshold;

[0029] If any pixel value in the temperature grayscale image is less than a pixel threshold, pixel conversion is performed on the pixel value according to a preset pixel value.

[0030] Furthermore, before inputting the binarized temperature grayscale image into a pre-trained temperature value recognition model for temperature value recognition, the method further includes:

[0031] Acquire historical infrared images, and crop the highest temperature value area and the lowest temperature value area in each historical infrared image to obtain a temperature value image;

[0032] Constructing a training set and a test set according to the temperature value image, and constructing the temperature value recognition model according to a DenseNet network and an LSTM network, wherein a random deactivation layer is provided in the LSTM network;

[0033] The temperature value recognition model is trained according to the training set until the temperature value recognition model converges, wherein the model training of the temperature value recognition model adopts stochastic gradient descent to perform parameter optimization.

[0034] Furthermore, after obtaining the infrared image of the power equipment to be identified, the method further includes:

[0035] Obtaining a device identification of the electric device to be identified, and matching the device identification with a pre-stored temperature and location query table;

[0036] If the device identifier successfully matches the temperature position lookup table, obtaining the total pixel value and the temperature display identifier of the device infrared image, and determining a cropping area based on the total pixel value, the temperature display identifier, and the device identifier;

[0037] In the infrared image of the device, the cropping area is cropped to obtain the temperature value area.

[0038] Another object of an embodiment of the present invention is to provide a temperature identification system for power equipment, the system comprising:

[0039] A keyword detection module is used to obtain an infrared image of the power equipment to be identified, and perform keyword detection on the infrared image of the equipment to obtain a keyword detection result;

[0040] a grayscale conversion module, configured to perform region cropping on the infrared image of the device according to the keyword detection result to obtain a temperature value region, and perform grayscale conversion on the temperature value region to obtain a temperature grayscale map;

[0041] A binarization module is used to perform pixel value conversion on the temperature grayscale image and perform binarization processing on the temperature grayscale image after pixel value conversion;

[0042] The temperature recognition module is used to input the binarized temperature grayscale image into a pre-trained temperature value recognition model to perform temperature value recognition, and obtain the maximum temperature value and the minimum temperature value of the power equipment to be identified.

[0043] Another object of an embodiment of the present invention is to provide a terminal device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above method when executing the computer program.

[0044] Another object of an embodiment of the present invention is to provide a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0045] The embodiment of the present invention performs keyword detection on the infrared image of the equipment to obtain the keyword detection result, and can automatically locate and crop the temperature value area based on the keyword detection result. The temperature value area facilitates the identification of the maximum temperature value and the minimum temperature value of the power equipment to be identified, thereby improving the efficiency of temperature identification of the power equipment. By performing pixel value conversion on the temperature grayscale image and binarizing the temperature grayscale image after pixel value conversion, the temperature value recognition model facilitates the temperature value recognition of the temperature value area. Based on the temperature value recognition model, the temperature value can be automatically recognized on the binarized temperature grayscale image to obtain the maximum temperature value and the minimum temperature value of the power equipment to be identified. There is no need to manually read and record the maximum temperature value and the minimum temperature value of the power equipment to be identified, thereby further improving the efficiency of temperature identification of the power equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 is a flow chart of a method for identifying temperature of electric power equipment provided by a first embodiment of the present invention;

[0047] Figures 2 to 6 is a schematic diagram of an infrared image of a device provided by the first embodiment of the present invention;

[0048] Figure 7 is a flow chart of a method for identifying temperature of electric power equipment provided by a second embodiment of the present invention;

[0049] Figure 8 2 is a schematic structural diagram of a temperature identification system for electric power equipment provided by a third embodiment of the present invention;

[0050] Figure 9 This is a flowchart of specific implementation steps of the power equipment temperature identification system provided by the third embodiment of the present invention;

[0051] Figure 10 This is a flow chart of temperature value positioning provided by the third embodiment of the present invention;

[0052] Figure 11 It is a schematic structural diagram of a terminal device provided in the fourth embodiment of the present invention. DETAILED DESCRIPTION

[0053] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0054] In order to illustrate the technical solution of the present invention, specific embodiments are provided below.

[0055] Example 1

[0056] See also Figure 1 , is a flow chart of a method for identifying the temperature of an electric power device provided by a first embodiment of the present invention. The method for identifying the temperature of an electric power device can be applied to any terminal device or system. The method for identifying the temperature of an electric power device comprises the following steps:

[0057] Step S10, obtaining an infrared image of the power equipment to be identified, and performing keyword detection on the infrared image of the equipment to obtain a keyword detection result;

[0058] Among them, an infrared image acquisition device can be used to capture images of the power equipment to be identified to obtain an infrared image of the equipment. By performing keyword detection on the infrared image of the equipment, a keyword detection result containing the positions of various keywords is obtained. The keyword can be set according to user needs. For example, the keyword can be set as text or a specified type of image, and the text includes text such as "highest" and "lowest". The specified type of image can be set as a color bar.

[0059] Optionally, in this step, after obtaining the infrared image of the power equipment to be identified, the step further includes:

[0060] Obtaining a device identification of the electric device to be identified, and matching the device identification with a pre-stored temperature location query table; wherein the temperature location query table stores a device identification of a specified device, the specified device being an electric device at a fixed location with a known temperature value;

[0061] If the device identifier successfully matches the temperature position lookup table, obtaining the total pixel value and the temperature display identifier of the device infrared image, and determining a cropping area based on the total pixel value, the temperature display identifier, and the device identifier;

[0062] In the infrared image of the device, the cropping area is cropped to obtain the temperature value area;

[0063] Among them, if the device identification successfully matches the temperature position query table, it is determined that the power equipment to be identified is a power equipment at a fixed position with a known temperature value. The cropping area is obtained by obtaining the total pixel value and temperature display identification of the device infrared image, and matching the total pixel value, temperature display identification and device identification with the pre-stored fixed position query table. The fixed position query table stores the correspondence between different total pixel values, temperature display identifications and device identifications and corresponding cropping areas. The temperature display identification is used to characterize the current temperature display mode on the infrared image of the device. The temperature display mode includes different position display relationships between the maximum temperature value and the minimum temperature value and the keyword "highest", the keyword "minimum" and the specified type of image "color bar" in the infrared image of the device;

[0064] For example, see Figure 2 When the total pixel value of the device's infrared image is 640*480, and the temperature display mode corresponding to the temperature display identifier is "the maximum and minimum temperature values ​​are located above and below the color bar", the cropped area is [[580,62,638,92], [580,415,638,445]], where the two cropped areas correspond to the highest and lowest temperature areas;

[0065] See also Figure 3 , when the total pixel value of the device's infrared image is 640*480, and the temperature display mode corresponding to the temperature display identifier is "the maximum temperature value and the minimum temperature value are located below the "highest" and "lowest" keywords", the cropping area is [[15,25,65,55],[80,25,140,55]]. Figure 4 As shown, the green box is the cropping area;

[0066] See also Figure 4 , when the total pixel value of the device's infrared image is 640*480, and the temperature display mode corresponding to the temperature display identifier is "the maximum temperature value and the minimum temperature value are located to the right of the keywords "highest" and "lowest", the cropping area is [[60,90,155,120], [60,125,155,155]].

[0067] See also Figure 5 , when the total pixel value of the device's infrared image is 696*480, and the temperature display mode corresponding to the temperature display mark is "the maximum temperature value and the minimum temperature value are located above and below the color bar", the cropping area is [[640,70,672,95],[640,365,672,390]];

[0068] See also Figure 6 , when the total pixel value of the device's infrared image is 800*480, and the temperature display mode corresponding to the temperature display identifier is "the maximum temperature value and the minimum temperature value are located above and below the color bar", the cropping area is [[750,175,800,210],[750,438,800,470]].

[0069] Step S20, performing region cropping on the device infrared image according to the keyword detection result to obtain a temperature value region, and performing grayscale conversion on the temperature value region to obtain a temperature grayscale map;

[0070] Among them, the infrared image of the device is regionally cropped through the cropping area in the keyword detection result to obtain the temperature value area containing the highest temperature value and the lowest temperature value image, and the grayscale transformation is performed on the temperature value area to achieve the image preprocessing effect of the temperature value area.

[0071] Step S30, pixel value conversion is performed on the temperature grayscale image, and the temperature grayscale image after pixel value conversion is binarized;

[0072] By performing pixel value conversion on the temperature grayscale image and binarizing the temperature grayscale image after pixel value conversion, the temperature value recognition model can conveniently recognize the temperature values in the temperature value region. Optionally, in this step, the pixel value conversion on the temperature grayscale image comprises:

[0073] The number of occurrences of each pixel value in the temperature grayscale image is obtained, and the average value between the pixel values corresponding to the maximum two occurrence numbers is calculated to obtain a pixel threshold. By obtaining the number of occurrences of each pixel value in the temperature grayscale image, the accuracy of pixel threshold calculation is improved based on the number of occurrences of each pixel value.

[0074] If any pixel value in the temperature grayscale image is less than the pixel threshold, the pixel value is converted according to a preset pixel value. The preset pixel value can be parameterized according to requirements.

[0075] Step S40, the temperature grayscale image after binarization is input into the pre-trained temperature value recognition model for temperature value recognition to obtain the highest temperature value and the lowest temperature value of the to-be-recognized power equipment.

[0076] Based on the temperature value recognition model, the temperature value recognition model can automatically recognize the temperature values of the temperature grayscale image after binarization to obtain the highest temperature value and the lowest temperature value of the to-be-recognized power equipment. The highest temperature value and the lowest temperature value of the to-be-recognized power equipment do not need to be read and recorded manually, and the power equipment temperature recognition efficiency is further improved.

[0077] Optionally, before the temperature grayscale image after binarization is input into the pre-trained temperature value recognition model for temperature value recognition, the method further comprises:

[0078] Obtain historical infrared images, and crop the highest temperature value region and the lowest temperature value region in each historical infrared image to obtain a temperature value image.

[0079] Construct a training set and a test set according to the temperature value image, and construct the temperature value recognition model according to a DenseNet network and an LSTM network, wherein a random inactivation layer is arranged in the LSTM network.

[0080] Model training is performed on the temperature value recognition model according to the training set until the temperature value recognition model converges, wherein the model training of the temperature value recognition model uses a stochastic gradient descent for parameter optimization.

[0081] Specifically, in this step, the temperature grayscale image is identified based on the CRNN network + CTC network, where:

[0082] Dataset: 4,000 historical infrared images were randomly selected. Two regions with the highest and lowest temperature values ​​were cropped from each image, resulting in a total of 8,000 temperature images. Each temperature image contained 11 characters, and the training and test sets were divided into a ratio of 8:2.

[0083] CRNN+CTC network training: The CNN network uses the DenseNet network, and the RNN uses a two-layer LSTM network with 64 neurons in each layer. Each LSTM layer is followed by a dropout layer with dropout_rate = 0.4 and network parameter batch_size = 32. Stochastic gradient descent (SGD) algorithm is used for optimization. The initial learning rate is 1e-3, momentum = 0.9, the maximum number of iterations is set to 300, and an early stopping strategy is used to prevent model overfitting.

[0084] In this embodiment, keyword detection is performed on the infrared image of the equipment to obtain keyword detection results. Based on the keyword detection results, the temperature value area can be automatically located and cropped. Based on the temperature value area, the maximum temperature value and the minimum temperature value of the power equipment to be identified are convenient for identification, thereby improving the efficiency of temperature identification of the power equipment. By performing pixel value conversion on the temperature grayscale image and binarizing the temperature grayscale image after pixel value conversion, the temperature value recognition model is convenient for identifying the temperature value of the temperature value area. Based on the temperature value recognition model, the temperature value can be automatically identified on the binarized temperature grayscale image to obtain the maximum temperature value and the minimum temperature value of the power equipment to be identified. There is no need to manually read and record the maximum temperature value and the minimum temperature value of the power equipment to be identified, thereby further improving the efficiency of temperature identification of the power equipment.

[0085] Example 2

[0086] See also Figure 7 , is a flow chart of a method for identifying the temperature of electric power equipment provided by a second embodiment of the present invention. This embodiment is used to further refine step S10 in the first embodiment, including the steps of:

[0087] Step S11, performing color bar detection on the infrared image of the device according to the pre-trained target detection model;

[0088] Among them, 4,000 infrared images of power equipment with known temperature values ​​at fixed locations were randomly selected as the data set, and the training set and test set were divided into two sets in a ratio of 8:2;

[0089] Training of the object detection model: The object detection model uses the YOLOv4 model, and the feature extraction network uses CSPdarknet53. The ImageNet pre-trained weights are used as the initial training parameters. Backpropagation and random gradient descent are used for end-to-end network training. For the first 50 iterations, the batch size is set to 8 and the learning rate is 1e-3. After 50 iterations, the batch size is set to 4 and the learning rate is 1e-4. The maximum number of iterations is 500, and an early stopping strategy is used to prevent model overfitting. The training is completed until the object detection model converges, resulting in the pre-trained object detection model.

[0090] Step S12: if the color mark bar is detected in the infrared image of the device, obtaining the location area of ​​the color mark bar and obtaining the keyword detection result;

[0091] If a color bar is detected in the infrared image of the device, the keyword detection result is obtained by obtaining the location area of ​​the color bar;

[0092] Step S13: If the color bar is not detected in the device infrared image, performing preset character detection on the device infrared image according to the target detection model;

[0093] The preset characters include a first character and a second character. In this step, the first character is set to "highest" and the second character is set to "lowest".

[0094] Step S14: if the preset character is detected in the infrared image of the device, obtaining the position area of ​​the first character and the position area of ​​the second character to obtain the keyword detection result;

[0095] Among them, if a preset character is detected in the device infrared image, the "highest" position area and the "lowest" position area are obtained to obtain the keyword detection result. Optionally, in this embodiment, for step S20, the device infrared image is regionally cropped according to the keyword detection result to obtain the temperature value area, including:

[0096] If the keyword detection result includes the location area of ​​the color mark bar, then the area above and below the location area of ​​the color mark bar at a first preset distance is determined as the cropping area; wherein the first preset distance can be set as a numerical value according to requirements;

[0097] If the keyword detection result includes the position area of ​​the first character and the position area of ​​the second character, obtaining the position relationship between the first character and the second character;

[0098] If the first character and the second character are in a horizontal relationship, an area at a second preset distance below the position area of ​​the first character and the position area of ​​the second character is determined as the cropping area; wherein the second preset distance can be set as required;

[0099] If the first character and the second character are in a vertical relationship, an area at a third preset distance to the right of the position area of ​​the first character and the position area of ​​the second character is determined as the cropping area; wherein the third preset distance can be set as required;

[0100] In the infrared image of the device, the cropping area is cropped to obtain the temperature value area.

[0101] Furthermore, in this embodiment, with respect to step S10, after performing keyword detection on the infrared image of the device and obtaining the keyword detection result, the method further includes:

[0102] If the color bar and the preset characters are not detected in the infrared image of the device, the device type and voltage level of the power device to be identified are matched with historical infrared data to obtain matching data;

[0103] The device type and voltage level of the power equipment to be identified are matched with the historical infrared data to query the historical infrared images corresponding to the power equipment with the same device type and voltage level as the power equipment to be identified;

[0104] Obtaining shooting year and month information of the infrared image of the device, and filtering the matching data according to the shooting year and month information to obtain first filtered data;

[0105] The matching data is filtered by shooting year and month information to further filter out historical infrared images shot by the device infrared image in the same year and / or month to obtain first filtered data;

[0106] Acquire the device address and substation name of the power device to be identified, and filter the first filtering data according to the device address and the substation name to obtain second filtering data;

[0107] The first screening data is screened by the device address and the substation name to further screen out historical infrared images with the same device address and the same substation name as the power device to be identified, thereby obtaining the second screening data;

[0108] Obtaining an image capture time of the device infrared image, and respectively calculating a time difference between a historical capture time of each historical infrared image in the second filtered data and the image capture time;

[0109] The historical infrared image corresponding to the minimum time difference is determined as the target infrared image, and the highest temperature value and the lowest temperature value of the target infrared image are determined as the highest temperature value and the lowest temperature value of the power equipment to be identified.

[0110] In this embodiment, the pre-trained target detection model is used to perform color mark bar detection on the device infrared image to detect whether there is a color mark bar in the device infrared image. If the color mark bar is detected in the device infrared image, the keyword detection result is obtained by obtaining the position area of ​​the color mark bar. If the color mark bar is not detected in the device infrared image, the device infrared image is subjected to preset character detection through the target detection model to detect whether there is a preset character in the device infrared image. If the preset character is detected in the device infrared image, the "highest" position area and the "lowest" position area are obtained to obtain the keyword detection result.

[0111] Example 3

[0112] See also Figure 8 , is a schematic diagram of the structure of a power equipment temperature recognition system 100 provided in a third embodiment of the present invention, comprising: a keyword detection module 10, a grayscale conversion module 11, a binarization module 12, and a temperature recognition module 13, wherein:

[0113] The keyword detection module 10 is used to obtain an infrared image of the power equipment to be identified, and perform keyword detection on the infrared image to obtain a keyword detection result.

[0114] The keyword detection module 10 is further configured to: perform color bar detection on the infrared image of the device according to the pre-trained target detection model;

[0115] If the color mark bar is detected in the infrared image of the device, obtaining the location area of ​​the color mark bar and obtaining the keyword detection result;

[0116] If the color mark bar is not detected in the device infrared image, performing preset character detection on the device infrared image according to the target detection model, the preset characters including a first character and a second character;

[0117] If the preset character is detected in the infrared image of the device, the position area of ​​the first character and the position area of ​​the second character are acquired to obtain the keyword detection result.

[0118] Optionally, the keyword detection module 10 is further configured to: if the color bar and the preset characters are not detected in the infrared image of the device, match the device type and voltage level of the power device to be identified with historical infrared data to obtain matching data;

[0119] Obtaining shooting year and month information of the infrared image of the device, and filtering the matching data according to the shooting year and month information to obtain first filtered data;

[0120] Acquire the device address and substation name of the power device to be identified, and filter the first filtering data according to the device address and the substation name to obtain second filtering data;

[0121] Obtaining an image capture time of the device infrared image, and respectively calculating a time difference between a historical capture time of each historical infrared image in the second filtered data and the image capture time;

[0122] The historical infrared image corresponding to the minimum time difference is determined as the target infrared image, and the highest temperature value and the lowest temperature value of the target infrared image are determined as the highest temperature value and the lowest temperature value of the power equipment to be identified.

[0123] Furthermore, the keyword detection module 10 is further configured to: obtain the device identification of the power device to be identified, and match the device identification with a pre-stored temperature location query table;

[0124] If the device identifier successfully matches the temperature position lookup table, obtaining the total pixel value and the temperature display identifier of the device infrared image, and determining a cropping area based on the total pixel value, the temperature display identifier, and the device identifier;

[0125] In the infrared image of the device, the cropping area is cropped to obtain the temperature value area.

[0126] The grayscale conversion module 11 is used to perform region cropping on the device infrared image according to the keyword detection result to obtain a temperature value region, and perform grayscale conversion on the temperature value region to obtain a temperature grayscale map.

[0127] The grayscale conversion module 11 is further configured to: if the keyword detection result includes the location area of ​​the color mark bar, determine the area above and below the location area of ​​the color mark bar at a first preset distance as the cropping area;

[0128] If the keyword detection result includes the position area of ​​the first character and the position area of ​​the second character, obtaining the position relationship between the first character and the second character;

[0129] If the first character and the second character are in a horizontal relationship, an area at a second preset distance below the position area of ​​the first character and the position area of ​​the second character is determined as the cropping area;

[0130] If the first character and the second character are in a vertical relationship, an area at a third preset distance to the right of the position area of ​​the first character and the position area of ​​the second character is determined as the cropping area;

[0131] In the infrared image of the device, the cropping area is cropped to obtain the temperature value area.

[0132] The binarization module 12 is configured to perform pixel value conversion on the temperature grayscale image and perform binarization processing on the temperature grayscale image after pixel value conversion.

[0133] The binarization module 12 is further configured to: obtain the number of occurrences of each pixel value in the temperature grayscale image, and calculate the average value between the two pixel values ​​corresponding to the maximum number of occurrences to obtain a pixel threshold;

[0134] If any pixel value in the temperature grayscale image is less than a pixel threshold, pixel conversion is performed on the pixel value according to a preset pixel value.

[0135] The temperature recognition module 13 is used to input the binarized temperature grayscale image into a pre-trained temperature value recognition model to perform temperature value recognition, and obtain the maximum temperature value and the minimum temperature value of the power equipment to be recognized.

[0136] The temperature recognition module 13 is further configured to: obtain historical infrared images, and crop the highest temperature value region and the lowest temperature value region in each historical infrared image to obtain a temperature value image;

[0137] Constructing a training set and a test set according to the temperature value image, and constructing the temperature value recognition model according to a DenseNet network and an LSTM network, wherein a random deactivation layer is provided in the LSTM network;

[0138] The temperature value recognition model is trained according to the training set until the temperature value recognition model converges, wherein the model training of the temperature value recognition model adopts stochastic gradient descent to perform parameter optimization.

[0139] Optionally, in this embodiment, please refer to Figure 9 The specific implementation steps of the power equipment temperature recognition system 100 include: positioning and cutting the temperature value area, using the adaptive threshold method to perform infrared image preprocessing and temperature value recognition on the cut temperature value area:

[0140] In the steps of locating and cropping the temperature value region, the temperature value region is located and cropped for the device infrared image to obtain the temperature value region;

[0141] In the infrared image preprocessing step, the temperature value area is first grayscale transformed to obtain a temperature grayscale map, the temperature grayscale map is pixel-wise converted, and the temperature grayscale map is converted from white text on a black background to black text on a white background. The temperature grayscale map after pixel value conversion is binarized using the OTSU algorithm to obtain a binary image.

[0142] Further, see Figure 10 , which is a flow chart of temperature value positioning provided in this embodiment, wherein:

[0143] If the fixed location of the temperature value is known, the area with the highest and lowest temperature values ​​can be directly located according to the location displayed by the temperature value information;

[0144] If the fixed location of the unknown temperature value cannot be directly obtained, the target detection model (YOLOv4) is used to detect the location of the color bar, "maximum", "minimum" and other keywords in the device's infrared image. Then, based on the display relationship between the temperature value and the color bar, "maximum", "minimum" and other keywords, the location of the highest and lowest temperature values ​​is obtained to obtain the temperature value area.

[0145] Step 1: If the fixed position of the temperature value is known, directly locate the temperature value area according to the pixel size of the device's infrared image and the temperature display method;

[0146] Step 2: The highest and lowest temperatures are above and below the color bar respectively. The areas with the highest and lowest values ​​are located by the position of the color bar to obtain the temperature value area.

[0147] Step 3: If the "highest" and "lowest" keywords are in a horizontal position, the lower area is the temperature value area; if the "highest" and "lowest" keywords are in a vertical position, the right area is the temperature value area;

[0148] Step 4: For the device infrared images where the highest and lowest temperature values ​​cannot be obtained based on the relevant information, match the "shooting time [year, month, day, hour, minute, second], city, substation name, device type, voltage level" information of the device infrared image with the fields in the database table. The highest and lowest temperature values ​​of the historical infrared image found last are used as the highest and lowest temperature values ​​of the device infrared image. The search algorithm is as follows:

[0149] First, find all data with the same field values ​​based on the "Device Type" and "Voltage Level" of the device infrared image;

[0150] Secondly, filter all the data of the "month" when the infrared image of the device was taken. If the "year" is the same, further filter out the data with the same year;

[0151] Then, filter by "City" and "Substation Name". If there is data for the same city in the table, further filter by "Substation Name". If there is no data for the same city in the table, search for data for adjacent cities.

[0152] Finally, according to the "shooting time" of the device infrared image, the historical infrared image with the shortest time difference with the device infrared image is searched in the table, and the highest and lowest temperature values ​​of the device infrared image found are used as the highest and lowest temperature values ​​of the device infrared image.

[0153] For the above steps 1, 2, and 3, the area where the temperature value is located in the infrared image is cropped according to the position information to obtain the temperature value area.

[0154] In this embodiment, keyword detection is performed on the infrared image of the equipment to obtain keyword detection results. Based on the keyword detection results, the temperature value area can be automatically located and cropped. Based on the temperature value area, the maximum temperature value and the minimum temperature value of the power equipment to be identified are convenient for identification, thereby improving the efficiency of temperature identification of the power equipment. By performing pixel value conversion on the temperature grayscale image and binarizing the temperature grayscale image after pixel value conversion, the temperature value recognition model is convenient for identifying the temperature value of the temperature value area. Based on the temperature value recognition model, the temperature value can be automatically identified on the binarized temperature grayscale image to obtain the maximum temperature value and the minimum temperature value of the power equipment to be identified. There is no need to manually read and record the maximum temperature value and the minimum temperature value of the power equipment to be identified, thereby further improving the efficiency of temperature identification of the power equipment.

[0155] Example 4

[0156] Figure 11 This is a structural block diagram of a terminal device 2 provided in the fourth embodiment of the present application. Figure 11 As shown, the terminal device 2 of this embodiment includes: a processor 20, a memory 21, and a computer program 22 stored in the memory 21 and executable on the processor 20, such as a program for a method for identifying the temperature of an electric power device. When the processor 20 executes the computer program 22, the steps of each embodiment of the method for identifying the temperature of an electric power device described above are implemented.

[0157] Exemplarily, the computer program 22 may be divided into one or more modules, which are stored in the memory 21 and executed by the processor 20 to implement the present application. The one or more modules may be a series of computer program instruction segments capable of performing specific functions, which are used to describe the execution process of the computer program 22 in the terminal device 2. The terminal device may include, but is not limited to, a processor 20 and a memory 21.

[0158] The processor 20 may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0159] The memory 21 may be an internal storage unit of the terminal device 2, such as a hard disk or memory of the terminal device 2. The memory 21 may also be an external storage device of the terminal device 2, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the terminal device 2. Furthermore, the memory 21 may include both an internal storage unit of the terminal device 2 and an external storage device. The memory 21 is used to store the computer program and other programs and data required by the terminal device. The memory 21 may also be used to temporarily store data that has been output or is about to be output.

[0160] In addition, the functional modules in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0161] If the integrated module is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Among them, the computer-readable storage medium can be non-volatile or volatile. Based on this understanding, the present application implements all or part of the process in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form, etc. The computer-readable storage medium may include: any entity or device that can carry computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content contained in computer-readable storage media can be appropriately increased or decreased according to the requirements of legislation and patent practices in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practices, computer-readable storage media do not include electrical carrier signals and telecommunications signals.

[0162] The above-described embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present application, and should all be included in the scope of protection of the present application.

Claims

1. A method for identifying the temperature of an electric power device, characterized in that: The method comprises: Acquire an infrared image of the power equipment to be identified, and perform keyword detection on the infrared image of the equipment to obtain a keyword detection result; Performing region cropping on the infrared image of the device according to the keyword detection result to obtain a temperature value region, and performing grayscale transformation on the temperature value region to obtain a temperature grayscale map; Performing pixel value conversion on the temperature grayscale image, and performing binarization processing on the temperature grayscale image after pixel value conversion; Inputting the binarized temperature grayscale image into a pre-trained temperature value recognition model to perform temperature value recognition, thereby obtaining the maximum temperature value and the minimum temperature value of the power equipment to be recognized; The performing keyword detection on the infrared image of the device includes: Performing color bar detection on the infrared image of the device according to the pre-trained target detection model; If the color mark bar is detected in the infrared image of the device, obtaining the location area of ​​the color mark bar and obtaining the keyword detection result; If the color mark bar is not detected in the device infrared image, performing preset character detection on the device infrared image according to the target detection model, the preset characters including a first character and a second character; If the preset character is detected in the infrared image of the device, obtaining the position area of ​​the first character and the position area of ​​the second character to obtain the keyword detection result; The step of performing region cropping on the device infrared image according to the keyword detection result to obtain a temperature value region includes: If the keyword detection result includes the location area of ​​the color mark bar, determining the area above and below the location area of ​​the color mark bar by a first preset distance as the cropping area; If the keyword detection result includes the position area of ​​the first character and the position area of ​​the second character, obtaining the position relationship between the first character and the second character; If the first character and the second character are in a horizontal relationship, an area at a second preset distance below the position area of ​​the first character and the position area of ​​the second character is determined as the cropping area; If the first character and the second character are in a vertical relationship, an area at a third preset distance to the right of the position area of ​​the first character and the position area of ​​the second character is determined as the cropping area; In the infrared image of the device, the cropping area is cropped to obtain the temperature value area; After performing keyword detection on the infrared image of the device and obtaining the keyword detection result, the method further includes: If the color bar and the preset characters are not detected in the infrared image of the device, the device type and voltage level of the power device to be identified are matched with historical infrared data to obtain matching data; Obtaining shooting year and month information of the infrared image of the device, and filtering the matching data according to the shooting year and month information to obtain first filtered data; Acquire the device address and substation name of the electric power device to be identified, and filter the first filtering data according to the device address and the substation name to obtain second filtering data; Obtaining an image capture time of the device infrared image, and respectively calculating a time difference between a historical capture time of each historical infrared image in the second filtered data and the image capture time; Determine the historical infrared image corresponding to the minimum time difference as the target infrared image, and determine the maximum temperature value and the minimum temperature value of the target infrared image as the maximum temperature value and the minimum temperature value of the power equipment to be identified; Before inputting the binarized temperature grayscale image into a pre-trained temperature value recognition model for temperature value recognition, the method further includes: Acquire historical infrared images, and crop the highest temperature value area and the lowest temperature value area in each historical infrared image to obtain a temperature value image; Constructing a training set and a test set according to the temperature value image, and constructing the temperature value recognition model according to a DenseNet network and an LSTM network, wherein a random deactivation layer is provided in the LSTM network; The temperature value recognition model is trained according to the training set until the temperature value recognition model converges, wherein the model training of the temperature value recognition model adopts stochastic gradient descent to perform parameter optimization.

2. The method for identifying the temperature of electric power equipment according to claim 1, wherein: The pixel value conversion of the temperature grayscale image includes: Obtain the number of occurrences of each pixel value in the temperature grayscale image respectively, and calculate the average value between the two pixel values ​​corresponding to the maximum number of occurrences to obtain a pixel threshold; If any pixel value in the temperature grayscale image is less than a pixel threshold, pixel conversion is performed on the pixel value according to a preset pixel value.

3. The method for identifying the temperature of electric power equipment according to claim 1, wherein: After obtaining the infrared image of the power equipment to be identified, the method further includes: Obtaining a device identification of the electric device to be identified, and matching the device identification with a pre-stored temperature and location query table; If the device identifier successfully matches the temperature position lookup table, obtaining the total pixel value and the temperature display identifier of the device infrared image, and determining a cropping area based on the total pixel value, the temperature display identifier, and the device identifier; In the infrared image of the device, the cropping area is cropped to obtain the temperature value area.

4. A power equipment temperature identification system, used to implement a power equipment temperature identification method according to any one of claims 1 to 3, characterized in that: include: A keyword detection module is used to obtain an infrared image of the power equipment to be identified, and perform keyword detection on the infrared image of the equipment to obtain a keyword detection result; a grayscale conversion module, configured to perform region cropping on the infrared image of the device according to the keyword detection result to obtain a temperature value region, and perform grayscale conversion on the temperature value region to obtain a temperature grayscale map; A binarization module is used to perform pixel value conversion on the temperature grayscale image and perform binarization processing on the temperature grayscale image after pixel value conversion; The temperature recognition module is used to input the binarized temperature grayscale image into a pre-trained temperature value recognition model to perform temperature value recognition, and obtain the maximum temperature value and the minimum temperature value of the power equipment to be identified.

5. A terminal device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 3 are implemented.

6. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 3 are implemented.

Images