CN115082318B - A super-resolution reconstruction method for infrared images of electrical equipment - Google Patents

Abstract

The invention discloses a super-resolution reconstruction method of an infrared image of an electrical device, which consists of a countermeasure network and a generation network, wherein the generation network is used for generating the super-resolution reconstruction image of the infrared image of the electrical device, and the countermeasure network is used for judging whether the infrared image of the electrical device with high resolution is a generated image or an original image with high resolution. The invention improves ESRGAN defects, constructs an improved batch standardization module and a new feature extraction sub-module, constructs a feature extraction network for generating an countermeasure network by using the two modules, improves the feature extraction capability of the generated network, and further improves the quality of the reconstructed image; in addition, an improved batch standardization module is introduced into the ESRGAN countermeasure network, so that the discrimination capability of the countermeasure network is improved, the infrared image reconstruction capability of the power equipment of the generation network is improved, and the generation network can reconstruct high-quality high-resolution infrared images of the power equipment from low-resolution infrared images of the power equipment.

Description

A super-resolution reconstruction method for infrared images of electrical equipment

Technical Field

The invention belongs to the technical field of infrared image enhancement and relates to a super-resolution reconstruction method for infrared images of electrical equipment.

Background technique

The failure of electrical equipment is often manifested by the temperature change of the fault part of the equipment. Therefore, the infrared imager is used to collect infrared images of electrical equipment and analyze the infrared images, which can realize contactless fault detection of power equipment. Although thermal imagers have been widely used in power equipment status monitoring, due to the high cost of high-resolution infrared thermal imagers, most power companies use ordinary infrared thermal imagers, and the resolution of the infrared images of electrical equipment collected is low, which directly affects the accuracy of equipment status detection and fault diagnosis based on infrared images of power equipment. There are two ways to improve the resolution of infrared images of electrical equipment: one is to improve the hardware level of infrared thermal imagers, thereby improving the resolution of infrared images. This method is difficult and costly; the other is to reconstruct low-resolution images into high-resolution images based on the existing infrared thermal imagers and the infrared image super-resolution reconstruction method. This method is low-cost.

The existing super-resolution reconstruction of single infrared images of electrical equipment can be divided into three categories: super-resolution reconstruction methods of electrical equipment infrared images based on difference, super-resolution reconstruction methods of electrical equipment infrared images based on reconstruction models, and super-resolution reconstruction methods of electrical equipment infrared images based on learning. The super-resolution reconstruction method of electrical equipment infrared images based on interpolation is simple and easy to implement, and can obtain a smooth reconstructed image, but some detail information is seriously lost, the visual quality of the reconstructed image is poor, and the texture information is not obvious. The super-resolution reconstruction method of electrical equipment infrared images based on reconstruction models uses prior information to reconstruct infrared images. The prior information in different environments is different, which affects the quality of the reconstructed image. The super-resolution reconstruction method of electrical equipment infrared images based on learning uses a large amount of training data to enable the model to learn a certain correspondence between high-resolution electrical equipment infrared images and corresponding low-resolution electrical equipment infrared images, and then uses the trained model to reconstruct high-resolution infrared images from low-resolution electrical equipment infrared images, thereby realizing super-resolution reconstruction of power equipment infrared images.

The prior art super-resolution reconstruction method of infrared images of electrical equipment based on generative adversarial networks belongs to a super-resolution image reconstruction method based on learning. Although it can better reconstruct infrared images of electrical equipment than other methods, the contour edges of the reconstructed image are not clear enough, which affects the electrical equipment status monitoring and fault diagnosis based on infrared images of electrical equipment. Therefore, it is necessary to improve the performance of the super-resolution reconstruction method of infrared images of electrical equipment based on generative adversarial networks, and thus improve the quality of the reconstructed infrared images of electrical equipment.

Summary of the invention

The technical problem to be solved by the present invention is to overcome the shortcomings of the prior art and provide a method for super-resolution reconstruction of infrared images of electrical equipment to improve the quality of reconstructed infrared images of electrical equipment.

The solution to the technical problem of the present invention is: a method for super-resolution reconstruction of infrared images of electrical equipment, characterized in that it consists of two parts: an adversarial network and a generative network. The generative network is used to generate a super-resolution reconstructed image of the infrared image of the electrical equipment, and the adversarial network is used to determine whether the high-resolution infrared image of the electrical equipment is a generated image or an original high-resolution image. Through the game between the generative network and the adversarial network, the image generation ability of the generative network is improved. The specific steps are as follows:

Step 1, build the dataset:

① Use high-resolution infrared imagers to collect high-resolution infrared images of electrical equipment;

② Using the degradation function including isotropic Gaussian blur, anisotropic Gaussian blur, downsampling, 3D Gaussian noise, imager noise and JPEG noise, the degradation process of the infrared image of real electrical equipment is fitted;

③ Perform degradation processing on the collected high-resolution infrared imaging images to obtain the corresponding low-resolution electrical equipment infrared images; the obtained degraded low-resolution electrical equipment infrared images and the corresponding high-resolution electrical equipment infrared images form an image pair, and several image pairs constitute a data set. The data set is divided into two parts, one as a training data set and the other as a test data set;

Step 2: Build an improved batch normalization module:

Based on the conventional batch normalization module, a feature pixel standard variance adjustment module is designed to reduce the impact of the conventional normalization module on the feature pixel standard variance and improve the quality of the reconstructed image.

Step 3: Build the final feature extraction submodule:

Construct the final feature extraction submodule to extract the features of infrared images of electrical equipment;

Step 4: Construct the generation network in the electrical equipment infrared image super-resolution reconstruction network:

Using the improved batch normalization module constructed in step 2 and the final feature extraction submodule constructed in step 3, combined with the ideas of dense connection network and residual network, on the basis of ESRGAN network, the improved ESRGAN network generation network is constructed;

Step 5: Construct an adversarial network in the electrical equipment infrared image super-resolution reconstruction network:

Introduce the module constructed in step 2 into the adversarial network of the ESRGAN network to construct an improved adversarial network of the ESRGAN network;

Step 6: Train the electrical equipment infrared image super-resolution reconstruction network:

Using the training data set constructed in step 1 to train the generative network in the electrical equipment infrared image super-resolution reconstruction network constructed in step 4 and the adversarial network in the electrical equipment infrared image super-resolution reconstruction network constructed in step 5;

Step 7: Network model testing and evaluation:

Input the low-resolution electrical equipment infrared image in the test data set constructed in step 1 into the trained generation network in step 6, and output the corresponding reconstructed electrical equipment infrared super-resolution image; calculate the peak signal-to-noise ratio of the reconstructed electrical equipment infrared super-resolution image, and evaluate the natural image quality. If the peak signal-to-noise ratio and the natural image quality evaluation meet the actual application requirements, execute step 9, otherwise execute step 8;

Step 8: Model parameter adjustment:

Adjust the model parameters of the electrical equipment infrared image super-resolution reconstruction network constructed by steps 4 and 5, and return to step 6 to retrain;

Step 9: Model Application:

The generation network that meets the practical application requirements obtained in step 7 is applied to the super-resolution reconstruction of infrared images of electrical equipment, and a high-resolution infrared image of electrical equipment is reconstructed from the collected low-resolution infrared image of electrical equipment.

Furthermore, 80% of the data set in step 1 is used as a training data set, and 20% of the data set is used as a test data set.

Furthermore, the improved batch normalization module constructed in step 2 is expressed as:

The improved batch normalization module consists of two branches, the first branch is a conventional batch normalization module, and the second branch is composed of a standard deviation function (std()), a logarithmic function (log), a linear function (f) and an exponential function (exp()) in sequence; the first branch needs to perform batch normalization preprocessing on the input of the module and then use it as the output of the conventional batch normalization module, and the second branch needs to perform standard deviation function, logarithmic function, linear function and exponential function processing on the input of the module in sequence and then output it, and the output of the first branch is multiplied by the output of the second branch to obtain the final output of the improved batch normalization module.

Furthermore, the final feature extraction submodule constructed in step 3 is expressed as:

The final feature extraction submodule consists of the previous feature extraction submodule and the improved batch normalization module in step 2;

The early feature extraction submodule consists of two branches, the first branch is composed of a 1×1 convolution, a ReLU6 function, a 3×3 convolution, a ReLU6 function and a 1×1 convolution in sequence; the second branch is composed of a 1×1 convolution, a ReLU6 function and a residual network in sequence;

The residual network is composed of a 3×3 convolution and a jump connection. The output of the 3×3 convolution and the output of the jump connection are spliced in the channel dimension so that the number of channels of the input feature map and the output feature map of the second branch remain unchanged; the output of the first branch and the output of the second branch are superimposed on the feature map to obtain the output features of the previous feature extraction submodule;

The early feature extraction submodule is incorporated into the improved batch normalization module in step 2 as a network module to form the final feature extraction submodule.

Furthermore, the generative network in the electrical equipment infrared image super-resolution reconstruction network constructed in step 4 is expressed as:

Several modules are defined as follows:

IBN module: the improved batch normalization module constructed in step 2;

GM module: the final feature extraction submodule constructed in step 3;

IRDB module: residual dense module;

IRRDB module: a feature extraction module with residual structure;

GM_DB module: residual dense module based on the final feature extraction submodule;

GM_RDB module: a feature extraction module based on the residual dense module of the final feature extraction submodule;

The generation network is composed of a 3×3 convolution module, a feature extraction network with a multi-level residual structure, an upsampling module and two serially connected 3×3 convolution modules.

The feature extraction network of the multi-level residual structure is composed of a first feature extraction subnetwork, a second feature extraction subnetwork, a GM module and two different jump connections, and the first feature extraction subnetwork and the second feature extraction subnetwork are connected in series;

The first feature extraction subnetwork consists of 8 IRRDB modules;

Each IRRDB module consists of three first residual modules with the same structure, one IBN module and one characteristic coefficient;

Each first residual module constituting the IRRDB module consists of an IRDB module, 1 feature coefficient, and 1 skip connection;

Each IRDB module consists of 5 3×3 convolution modules, 4 LReLU functions, 1 feature coefficient, 1 skip connection and IBN module;

The second feature extraction subnetwork consists of 8 GM_DB modules and 8 GM_RDB modules;

Each GM_DB module consists of 4 GM modules, 4 LReLU functions, a 3×3 convolution module, 1 feature coefficient, 1 skip connection and an IBN module;

Each GM_RDB module consists of three second residual modules with the same structure, one IBN module, one feature coefficient and one skip connection;

Each second residual module that makes up GM_RDB consists of a GM_DB module, a feature coefficient and a skip connection; the output of the second feature extraction subnetwork is fused with the input of the first feature extraction subnetwork through a skip connection, and the fused features are input into the GM module. The output of the GM module is fused with the input of the first feature extraction subnetwork through a skip connection, and the fused input features are sequentially subjected to an upsampling and two 3×3 convolutional layers to obtain the reconstructed image of the infrared image of the electrical equipment.

Furthermore, the adversarial network in the electrical equipment infrared image super-resolution reconstruction network constructed in step 5 is expressed as:

On the basis of the adversarial network in ESRGN, the conventional batch normalization module in the adversarial network is replaced with the IBN module constructed in step 3 to form an adversarial network in the improved electrical equipment infrared image super-resolution reconstruction network; the adversarial network in the improved electrical equipment infrared image super-resolution reconstruction network is composed of 1 3×3 convolution module, 1 LReLU function, 4 submodules, 1 4×4 convolution module, 1 IBN module, 1 LReLU function, 1 widening layer, 1 fully connected layer, 1 LReLU function and 1 fully connected layer in sequence;

The number of channels of the four submodules are 128, 256, 512 and 512 respectively;

The four submodules have the same structure, which is composed of a 4×4 convolution, an IBN module, an LReLU function, a 3×3 convolution layer, an IBN module and an LReLU function in sequence.

The beneficial effects of the present invention are:

(1) The present invention constructs an improved batch normalization module (IBN module) to reduce the impact of conventional batch normalization on the variance of image feature pixels. At the same time, the improved batch normalization module is introduced into the generative network and the adversarial network to avoid gradient vanishing, improve the network training speed and generalization ability, and thus improve the quality of infrared image reconstruction of electrical equipment based on the generative adversarial network.

(2) The present invention constructs a final feature extraction submodule, and uses the final feature extraction submodule and the improved normalization module to construct a residual dense module (GM_DB module), and also uses the constructed GM_DB module and the improved batch normalization module to construct a GM_RDB module, and uses 8 GM_DB modules and 8 GM_RDB modules as part of the feature extraction network in the generation network;

(3) The present invention introduces an improved batch normalization module (IBN) into the RDB module in ESRGAN to obtain an improved RDB module (IRDB module). The IRRDB module is constructed using the IRDB module and the IBN module. The eight IRRDB modules are used as part of the feature extraction network in the generation network.

(4) The present invention also uses the GM network to further extract and fuse the output features of the network composed of 8 GM_DB modules, 8 GM_RDB modules and 8 IRRDB modules, and uses jump connections to fuse the extracted features with the input features of the network, thereby extracting more infrared image features of power equipment and improving the ability of the generation network to reconstruct infrared images of power equipment;

(5) The present invention also improves the adversarial network by using the improved batch normalization module (IBN module) constructed by the present invention, thereby improving the discrimination ability of the adversarial network, thereby improving the ability of the generation network to reconstruct infrared images of power equipment, and improving the quality of the reconstructed infrared images of power equipment.

The ESRGAN mentioned above is the paper ESRGAN: Enhanced Super-Resolution Generative Adversarial Networks[C] published by Wang X.T et al. in 2019. European Conference on Computervision, 2019: 63–79

BRIEF DESCRIPTION OF THE DRAWINGS

Fig. 1 is a flow chart of the present invention;

FIG2 is a network structure diagram of an improved batch normalization module proposed by the present invention, in which the network module is a module that needs to be subjected to batch normalization processing;

FIG3 is a network structure diagram of a feature extraction submodule constructed in the present invention;

FIG4 is a network structure diagram of an IRDB module constructed by the present invention and a network structure diagram of an IRRDB module composed of IRDB modules, wherein the IRRDB module is used to construct a feature extraction network in a generation network;

FIG5 is a network structure diagram of the GM_DB module and the GM_RDB module constructed by the present invention, both modules are used to construct a feature extraction network in a generation network;

6 is a diagram showing a generated network structure constructed based on an IBN module, a GM module, an IRRDB module, a GM_DB module and a GM_RDB module according to the present invention;

FIG. 7 is a diagram showing the structure of an adversarial network constructed by the present invention.

Detailed ways

The present invention will be further described below using the accompanying drawings and specific implementation methods.

Referring to FIG. 1 to FIG. 7 , a method for super-resolution reconstruction of infrared images of electrical equipment in this embodiment is used to improve the quality of reconstructed infrared images of power equipment. The method comprises a generation network and an adversarial network. The generation network is used to generate super-resolution reconstructed images of infrared images of electrical equipment. The adversarial network is used to determine whether a high-resolution infrared image of electrical equipment is a generated image or an original high-resolution image. The image generation capability of the generation network is improved through the game between the generation network and the adversarial network. The specific steps are as follows:

Step 1, build the dataset:

① Use high-resolution infrared imagers to collect high-resolution infrared images of electrical equipment;

② Using the degradation function including isotropic Gaussian blur, anisotropic Gaussian blur, downsampling, 3D Gaussian noise, imager noise and JPEG noise, the degradation process of the infrared image of real electrical equipment is fitted;

③ Perform degradation processing on the collected high-resolution infrared imaging images to obtain the corresponding low-resolution electrical equipment infrared images; the obtained degraded low-resolution electrical equipment infrared images and the corresponding high-resolution electrical equipment infrared images form an image pair, and several image pairs constitute a data set. The data set is divided into two parts, 80% of the data set is used as a training sample set, and 20% of the data set is used as a test set;

Step 2: Build an improved batch normalization module:

Based on the conventional batch normalization module, a feature pixel standard variance adjustment module is designed to reduce the impact of the conventional batch normalization module on the feature pixel standard variance and improve the quality of the reconstructed image.

Step 3: Build the final feature extraction submodule:

Constructing a final feature extraction submodule for extracting features of infrared images of electrical equipment to extract more effective features of infrared images of electrical equipment;

Step 4: Construct the generation network in the electrical equipment infrared image super-resolution reconstruction network:

Using the improved batch normalization module constructed in step 2 and the final feature extraction submodule constructed in step 3, combined with the ideas of dense connection network and residual network, on the basis of ESRGAN network, the improved ESRGAN network generation network is constructed;

Step 5: Construct an adversarial network in the electrical equipment infrared image super-resolution reconstruction network:

Introduce the module constructed in step 2 into the adversarial network of the ESRGAN network to construct an improved adversarial network of the ESRGAN network;

Step 6: Train the electrical equipment infrared image super-resolution reconstruction network:

Using the training data set constructed in step 1 to train the generative network in the electrical equipment infrared image super-resolution reconstruction network constructed in step 4 and the adversarial network in the electrical equipment infrared image super-resolution reconstruction network constructed in step 5;

Step 7: Network model testing and evaluation:

Input the low-resolution electrical equipment infrared image in the test data set constructed in step 1 into the generative network trained in step 6, and output the corresponding reconstructed electrical equipment infrared super-resolution image; calculate the peak signal-to-noise ratio of the reconstructed electrical equipment infrared super-resolution image, and evaluate the natural image quality. If the peak signal-to-noise ratio and the natural image quality evaluation meet the actual application requirements, execute step 9, otherwise execute step 8;

Step 8: Model parameter adjustment:

Adjust the model parameters of the electrical equipment infrared image super-resolution reconstruction network constructed by steps 4 and 5, and return to step 6 to retrain;

Step 9: Model Application:

The generation network that meets the practical application requirements obtained in step 7 is applied to the super-resolution reconstruction of infrared images of electrical equipment, and a high-resolution infrared image of electrical equipment is reconstructed from the collected low-resolution infrared image of electrical equipment.

The improved batch normalization module constructed in the steps is expressed as:

The improved batch normalization module is shown in Figure 2. The improved batch normalization module consists of two branches. The first branch is a conventional batch normalization module, and the second branch is composed of a standard deviation function (std()), a logarithmic function (log), a linear function (f) and an exponential function (exp()) in sequence. The first branch needs to perform batch normalization preprocessing on the input of the module and then use it as the output of the conventional batch normalization module. The second branch needs to perform standard deviation function, logarithmic function, linear function and exponential function processing on the input of the module in sequence and then output it. The output of the first branch is multiplied by the output of the second branch to obtain the final output of the improved batch normalization module.

The final feature extraction submodule constructed in step 3 is expressed as:

The construction of the final feature extraction submodule is shown in Figure 3. The final feature extraction submodule consists of the previous feature extraction submodule and the improved batch normalization module of step 2;

The early feature extraction submodule consists of two branches, the first branch is composed of a 1×1 convolution, a ReLU6 function, a 3×3 convolution, a ReLU6 function and a 1×1 convolution in sequence; the second branch is composed of a 1×1 convolution, a ReLU6 function and a residual network in sequence;

The residual network is composed of a 3×3 convolution and a jump connection. The output of the 3×3 convolution and the output of the jump connection are spliced in the channel dimension so that the number of channels of the input feature map and the output feature map of the second branch remain unchanged; the output of the first branch and the output of the second branch are superimposed on the feature map to obtain the output features of the previous feature extraction submodule;

The early feature extraction submodule is incorporated into the improved batch normalization module in step 2 as a network module to form the final feature extraction submodule.

The generation network in the electrical equipment infrared image super-resolution reconstruction network constructed in step 4 is expressed as:

The submodules that make up the generation network are shown in Figures 4 and 5, and the generation network is shown in Figure 6. Several modules are defined as follows:

IBN module: the improved batch normalization module constructed in step 2;

GM module: the final feature extraction submodule constructed in step 3;

IRDB module: residual dense module;

IRRDB module: a feature extraction module with residual structure;

GM_DB module: residual dense module based on the final feature extraction submodule;

GM_RDB module: a feature extraction module based on the residual dense module of the final feature extraction submodule;

The generation network is composed of a 3×3 convolution module, a feature extraction network with a multi-level residual structure, an upsampling module and two serially connected 3×3 convolution modules.

The feature extraction network of the multi-level residual structure is composed of a first feature extraction subnetwork, a second feature extraction subnetwork, a GM module and two different jump connections, and the first feature extraction subnetwork and the second feature extraction subnetwork are connected in series;

The first feature extraction subnetwork consists of 8 IRRDB modules;

Each IRRDB module consists of three first residual modules with the same structure, one IBN module and one characteristic coefficient;

Each first residual module constituting the IRRDB module consists of an IRDB module, 1 feature coefficient, and 1 skip connection;

Each IRDB module consists of 5 3×3 convolution modules, 4 LReLU functions, 1 feature coefficient, 1 skip connection and IBN module;

The second feature extraction subnetwork consists of 8 GM_DB modules and 8 GM_RDB modules;

Each GM_DB module consists of 4 GM modules, 4 LReLU functions, a 3×3 convolution module, 1 feature coefficient, 1 skip connection and an IBN module;

Each GM_RDB module consists of three second residual modules with the same structure, one IBN module, one feature coefficient and one skip connection;

Each second residual module that makes up GM_RDB consists of a GM_DB module, a feature coefficient and a skip connection; the output of the second feature extraction subnetwork is fused with the input of the first feature extraction subnetwork through a skip connection, and the fused features are input into the GM module. The output of the GM module is fused with the input of the first feature extraction subnetwork through a skip connection, and the fused input features are sequentially subjected to an upsampling and two 3×3 convolutional layers to obtain the reconstructed image of the infrared image of the electrical equipment.

The adversarial network in the electrical equipment infrared image super-resolution reconstruction network constructed in step 5 is expressed as:

The improved adversarial network is shown in FIG7 . On the basis of the adversarial network in ESRGN, the conventional batch normalization module in the adversarial network is replaced with the IBN module constructed in step 3 to form an improved adversarial network in the infrared image super-resolution reconstruction network of electrical equipment. The adversarial network in the improved infrared image super-resolution reconstruction network of electrical equipment is composed of 1 3×3 convolution module, 1 LReLU function, 4 submodules, 1 4×4 convolution module, 1 IBN module, 1 LReLU function, 1 widening layer, 1 fully connected layer, 1 LReLU function and 1 fully connected layer in sequence.

The number of channels of the four submodules are 128, 256, 512 and 512 respectively;

The four submodules have the same structure, which is composed of a 4×4 convolution, an IBN module, an LReLU function, a 3×3 convolution layer, an IBN module and an LReLU function in sequence.

The above description is only a preferred embodiment of the present invention and does not constitute any other form of limitation to the present invention. Any modification or equivalent change made based on the technical essence of the present invention still falls within the scope of protection required by the present invention.

Claims (1)

1. The super-resolution reconstruction method for the infrared image of the electrical equipment is characterized by comprising the following specific steps of:

Step 1, constructing a data set:

① Acquiring an infrared image of the electrical equipment with high resolution by using a high resolution infrared imager;

② Fitting a degradation process of an infrared image of real electrical equipment by using degradation functions comprising isotropic Gaussian blur, anisotropic Gaussian blur, downsampling, 3D Gaussian noise, imager noise and JPEG noise;

③ Performing degradation treatment on the acquired high-resolution infrared imaging image to obtain a corresponding low-resolution infrared image of the electrical equipment; the obtained degraded low-resolution infrared image of the electrical equipment and the corresponding high-resolution infrared image of the electrical equipment form image pairs, a plurality of image pairs form a data set, the data set is divided into two parts, one part is used as a training data set, and the other part is used as a test data set;

Step 2, constructing an improved batch standardization module:

Designing a characteristic pixel standard deviation adjusting module on the basis of a conventional batch standardization module;

Step 3, constructing a final feature extraction sub-module:

constructing a final feature extraction sub-module for extracting infrared image features of the electrical equipment;

step 4, constructing a generation network in an infrared image super-resolution reconstruction network of the electrical equipment:

Utilizing the improved batch standardization module constructed in the step 2 and the final feature extraction submodule constructed in the step 3 to construct a generation network in the improved infrared image super-resolution reconstruction network of the electrical equipment on the basis of the infrared image super-resolution reconstruction network of the electrical equipment;

Step 5, constructing an countermeasure network in the infrared image super-resolution reconstruction network of the electrical equipment:

Introducing the improved batch standardization module constructed in the step 2 into an countermeasure network of an infrared image super-resolution reconstruction network of the electrical equipment, and constructing the countermeasure network in the improved infrared image super-resolution reconstruction network of the electrical equipment;

step 6: training an infrared image super-resolution reconstruction network of electrical equipment:

Training the generation network in the electric equipment infrared image super-resolution reconstruction network constructed in the step 4 and the countermeasure network in the electric equipment infrared image super-resolution reconstruction network constructed in the step 5 by adopting the training data set constructed in the step 1;

Step 7: network model test and evaluation:

Inputting the low-resolution infrared images of the electrical equipment in the test data set constructed in the step 1 into the trained generation network in the step 6, and outputting corresponding reconstructed infrared super-resolution images of the electrical equipment; calculating the peak signal-to-noise ratio of the reconstructed infrared super-resolution image of the electrical equipment, and evaluating the quality of the natural image; if the peak signal-to-noise ratio and the natural image quality evaluation meet the actual application requirements, executing the step 9, otherwise executing the step 8;

Step 8: model parameter adjustment:

Adjusting the model parameters of the infrared image super-resolution reconstruction network of the electrical equipment constructed in the step 4 and the step 5, returning to the step 6, and retraining;

step 9: model application:

The generating network meeting the practical application requirements, which is obtained in the step 7, is applied to super-resolution reconstruction of the infrared image of the electrical equipment, and the high-resolution infrared image of the electrical equipment is reconstructed from the acquired low-resolution infrared image of the electrical equipment;

The build-up of the modified batch normalization module in step 2 is expressed as:

The improved batch standardization module consists of two branches, wherein the first branch is a conventional batch standardization module, and the second branch consists of a standard deviation function, a logarithmic function, a linear function and an exponential function in sequence; the first branch needs to perform batch standardization pretreatment on the input of the module and then is used as the output of a conventional batch standardization module, the second branch needs to sequentially perform standard deviation function, logarithmic function, linear function and exponential function treatment on the input of the module and then outputs the module, and the output of the first branch is multiplied with the output of the second branch to obtain the final output of the improved batch standardization module;

the final feature extraction submodule constructed in the step 3 is expressed as:

The final feature extraction sub-module consists of a front-stage feature extraction sub-module and an improved batch standardization module in the step 2;

The early feature extraction submodule consists of two branches, wherein the first branch consists of a convolution of 1 multiplied by 1, a ReLU6 function, a convolution of 3 multiplied by 3, a ReLU6 function and a convolution of 1 multiplied by 1 in sequence; the second branch consists of a convolution of 1×1, a ReLU6 function and a residual network in sequence;

The residual network consists of a 3X 3 convolution and jump connection, and the output of the 3X 3 convolution and the output of the jump connection are spliced in the channel dimension, so that the number of channels of the input characteristic diagram and the output characteristic diagram of the second branch is kept unchanged; superposing the output of the first branch and the output of the second branch to obtain the output characteristics of the early-stage characteristic extraction submodule;

Incorporating the early feature extraction submodule into the improved batch standardization module of the step 2 as a network module to form a final feature extraction submodule;

The generating network in the electrical equipment infrared image super-resolution reconstruction network constructed in the step 4 is expressed as follows:

The several modules are defined as follows:

IBN module: an improved batch standardization module constructed in the step 2;

GM module: a final feature extraction sub-module constructed in the step 3;

IRDB module: a residual error dense module;

IRRDB module: a feature extraction module having a residual structure;

Gm_db module: a residual dense module based on the final feature extraction sub-module;

Gm_rdb module: the feature extraction module is based on a residual error dense module of the final feature extraction sub-module;

the generating network sequentially comprises a 3X 3 convolution module, a characteristic extracting network of a multi-level residual error structure, an up-sampling module and two 3X 3 convolution modules which are connected in series;

The characteristic extraction network of the multi-level residual error structure consists of a first characteristic extraction sub-network, a second characteristic extraction sub-network, 1 GM module and two different jump connections, wherein the first characteristic extraction sub-network and the second characteristic extraction sub-network are connected in series;

The first feature extraction sub-network consists of 8 IRRDB modules;

Each IRRDB module consists of 3 first residual modules with the same structure, 1 IBN module and a characteristic coefficient;

Each first residual error module forming IRRDB modules consists of an IRDB module, 1 characteristic coefficient and 1 jump connection;

Each IRDB module consists of 53 multiplied by 3 convolution modules, 4 LReLU functions, 1 characteristic coefficient, 1 jump connection and an IBN module;

the second feature extraction sub-network consists of 8 GM_DB modules and 8 GM_RDB modules;

Each gm_db module consists of 4 GM modules, 4 LReLU functions, one 3 x 3 convolution module, 1 feature factor, 1 hopping connection, and IBN modules;

Each GM_RDB module consists of 3 second residual modules with the same structure, 1 IBN module, 1 characteristic coefficient and 1 jump connection;

Each second residual module forming the gm_rdb is formed by 1 gm_db module, 1 characteristic coefficient and 1 jump connection; the output of the second characteristic extraction sub-network and the input of the first characteristic extraction sub-network are fused through jump connection, the fused characteristics are input into the GM module, the output of the GM module and the input of the first characteristic extraction sub-network are fused through jump connection, and the fused input characteristics sequentially pass through an up-sampling layer and 2 3X 3 convolution layers to obtain an infrared image reconstruction image of the electrical equipment;

the construction of the countermeasure network in the electrical equipment infrared image super-resolution reconstruction network in the step 5 is represented as follows:

On the basis of an countermeasure network in an electrical equipment infrared image super-resolution reconstruction network, replacing a conventional batch standardization module in the countermeasure network with an IBN module constructed in the step 3 to form an improved countermeasure network in the electrical equipment infrared image super-resolution reconstruction network; the countermeasure network in the improved infrared image super-resolution reconstruction network of the electrical equipment sequentially comprises 13×3 convolution module, 1 LReLU function, 4 submodules, 14×4 convolution module, 1 IBN module, 1 LReLU function, 1 stretching layer, 1 full connection layer, 1 LReLU function and 1 full connection layer;

the number of channels of the 4 sub-modules is 128, 256, 512 and 512 respectively;

the 4 sub-modules have the same structure and sequentially consist of 14×4 convolution, one IBN module, one LReLU function, one 3×3 convolution layer, one IBN module and one LReLU function.