Multi-scale feature enhancement-based power grid line-colliding bird identification method and system

Abstract

The invention discloses a power grid line-colliding bird identification method and system based on multi-scale feature enhancement. The method comprises the following steps: constructing and training an RBOD rare bird target detection model; the trained RBOD valuable and rare bird target detection model is adopted to carry out power grid line collision bird identification; the RBOD valuable and rare bird target detection model comprises a feature extraction network, a feature fusion network and a detection network. The feature extraction network is sequentially composed of an SRFD module, a first CWDB module, a first DRFD module, a second CWDB module, a second DRFD module, a third CWDB module, a third DRFD module, a fourth CWDB module, an SPPF module and a PSA module. The feature fusion network is used for realizing effective integration of features of different scales. According to the method, the multi-scale feature processing capability is improved, the problem of gradient disappearance during training can be eliminated, and the rare bird image detection capability is effectively improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of transmission line inspection, and particularly relates to a method and system for identifying birds hitting power grids based on multi-scale feature enhancement. Background Art

[0002] Birds hitting transmission lines can trigger accidents such as short circuits, tripping, and even fires in transmission lines, causing huge economic losses. Traditional solutions such as physical devices like anti-bird spines and bird repellers can partially alleviate the problem, but they have limitations such as passive protection and interference with the ecological balance. Manual inspection is limited by efficiency and safety and is difficult to achieve full coverage. Against this background, the rise of artificial intelligence technology provides a new path for the collaborative optimization of power systems and bird protection.

[0003] In recent years, the application of artificial intelligence in bird protection in power systems has become increasingly widespread. Based on advanced artificial intelligence algorithms, operation and maintenance personnel can achieve precise monitoring of bird activities, and through high-precision and high-efficiency technical means, they can identify and locate bird activities near transmission lines in real time, thereby providing data support for subsequent protection measures. The intelligent system can achieve efficient identification and location of bird activities near transmission lines, providing strong technical support for the safe operation of power systems and bird protection. The wide application of this technology can not only reduce bird hitting incidents but also provide a scientific basis for the balance between ecological protection and industrial development. Summary of the Invention

[0004] In view of this, the purpose of the present invention is to provide a method and system for identifying birds hitting power grids based on multi-scale feature enhancement. The present invention improves on the YOLOv10 model as the basic model to construct the RBOD rare bird target detection model, which can accurately identify birds hitting power grids and provide strong technical support for the safe operation of power systems and the protection of rare birds.

[0005] To achieve the above purpose, the present invention provides the following technical solution: A method for identifying birds hitting power grids based on multi-scale feature enhancement, constructing and training the RBOD rare bird target detection model; using the trained RBOD rare bird target detection model to identify birds hitting power grids; The RBOD rare bird target detection model includes three parts: a feature extraction network, a feature fusion network, and a detection network. The feature extraction network is composed of an SRFD module, a first CWDB module, a first DRFD module, a second CWDB module, a second DRFD module, a third CWDB module, a third DRFD module, a fourth CWDB module, an SPPF module, and a PSA module in sequence; the feature fusion network is used to effectively integrate features of different scales; the DRFD module includes an upper branch and a lower branch: the lower branch of the DRFD module is pixel cutting downsampling, and in the upper branch of the DRFD module, first, grouped convolution is performed, and then through two branches, the first branch completes depthwise separable convolution and then uses the GELU activation function for feature enhancement, and the second branch performs max pooling downsampling; the output features of the two branches are concatenated and convolved with the output of the lower branch, and finally, the output of the DRFD module is obtained; The CWDB module is composed of a CBS module, a separation module, several WDBB modules, a fusion module, and another CBS module in sequence.

[0006] Further preferably, the processing process of the feature fusion network is as follows: the output feature F of the PSA module 3 After passing through the first upsampling module, it is fused with the output feature F of the third CWDB module 2 Through the first fusion module for fusion, the obtained fusion feature C 1 Is input into the fifth CWDB module for processing; the output feature C of the fifth CWDB module 2 After being processed by the upsampling module, it is fused with the output feature F of the second CWDB module 1 Through the second fusion module for fusion, the obtained fusion feature C 3 ; the fusion feature C 3 After being processed by the sixth CWDB module, the output feature C of the sixth CWDB module 4 After being processed by the fourth DRFD module, it is fused with the output feature C of the fifth CWDB module 2 Through the third fusion module for feature fusion, the obtained fusion feature C 5 , the fusion feature C 5 Is input into the seventh CWDB module for processing, and the output feature C of the seventh CWDB module 6 After being processed by the fifth DRFD module, it is fused with F 3 Through the fourth fusion module for fusion, the obtained fusion feature C 7 , the fusion feature C 7 Is input into the C2FCIB module for processing, the output feature C of the sixth CWDB module 4 , the output feature C of the seventh CWDB module 6 , the output feature C of the C2FCIB module 8As the input of the three detection modules of the detection network.

[0007] Further preferably, the WDBB module divides the input features into six branches for processing. The first branch is sequentially processed by a 1×1 ordinary convolution and a batch normalization layer; the second branch is sequentially processed by a 1×1 ordinary convolution, the first batch normalization layer, a k×k ordinary convolution, and the second batch normalization layer; the third branch is sequentially processed by a 1×1 ordinary convolution, the first batch normalization layer, an average pooling layer, and the second batch normalization layer; the fourth branch is sequentially processed by a k×k ordinary convolution and a batch normalization layer; the fifth branch is sequentially processed by a 1×k ordinary convolution and a batch normalization layer; the sixth branch is sequentially processed by a k×1 ordinary convolution and a batch normalization layer; the features processed by the six branches are subjected to convolutional summation, and the output features are obtained after passing through the activation function SiLU.

[0008] Further preferably, the PSA module operates as follows: The module is divided into seven parts, namely the first ordinary convolution, the multi-head self-attention module, the first feature splicing module, the feed-forward network, the second feature splicing module, the fusion module, and the second ordinary convolution. The output features of the first ordinary convolution are respectively input to the multi-head self-attention module, the first feature splicing module, and the fusion module. The features after splicing by the first feature splicing module are respectively input to the feed-forward network and the second feature splicing module. The features are fused by the fusion module and then input to the second ordinary convolution for convolutional processing to obtain the final output of the PSA module.

[0009] Further preferably, the processing process of the C2FCIB module is as follows: The image features are input to a CBS module for convolutional processing, then processed by a separation module, and then sequentially processed by several CIB modules. Finally, the features of the separation module and several CIB modules are fused through a fusion module, and the fused features are input to another CBS module for processing to obtain the output of the C2FCIB module.

[0010] Further preferably, the shallow feature extraction module is the SRFD module. The SRFD module includes three stages. The first stage uses ordinary convolution; the features output in the first stage enter the second stage. The upper branch of the second stage consists of grouped convolution and depthwise separable convolution, and the lower branch of the second stage is pixel cutting downsampling. The features of the upper branch and the lower branch are spliced and convolved and then enter the third stage; in the upper branch of the third stage, grouped convolution is first performed, and then depthwise separable convolution and max-pooling downsampling are respectively performed through two branches. The features output by depthwise separable convolution and max-pooling downsampling are spliced and convolved with the output of the lower branch of the third stage, and finally the output features of the SRFD module are obtained.

[0011] Further preferably, the detection module consists of two branches. The operation mode of the first branch is as follows: after the input features are convolved by the CBS module, they are sequentially processed by the MultiSEAM module and ordinary convolution to obtain the prediction loss. The operation mode of the second branch is as follows: after the input features are processed by two consecutive CBS modules, they are sequentially processed by the MultiSEAM module and ordinary convolution to obtain the classification loss.

[0012] Further preferably, according to the historical data of rare bird collision accidents on the power grid, rare bird images are collected through transmission line monitoring cameras and network resources. A lightweight image super-resolution reconstruction model is used to super-resolve the rare bird images to improve the image resolution. After the image reconstruction is completed, a dataset of rare birds that accidentally collide with the transmission line is constructed. The dataset of rare birds that accidentally collide with the transmission line is used to train the RBOD rare bird target detection model.

[0013] The present invention also provides a power grid crossing line bird recognition system based on multi-scale feature enhancement, including an image acquisition device and a power grid crossing line bird recognition device. The image acquisition device is used to collect bird images within a certain distance range of the transmission line. The power grid crossing line bird recognition device is built-in with an RBOD rare bird target detection model for bird recognition.

[0014] The present invention also provides an electronic device, including a memory and a processor. The memory stores computer-readable instructions. When the instructions are executed by the processor, the processor is caused to implement the above-mentioned power grid crossing line bird recognition method based on multi-scale feature enhancement.

[0015] The present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the above-mentioned power grid crossing line bird recognition method based on multi-scale feature enhancement is implemented.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: First, rare bird images are collected through transmission line monitoring cameras and network resources, and a lightweight image super-resolution reconstruction model is used to perform super-resolution reconstruction on the rare bird images, aiming to improve the image resolution, enhance the effectiveness of the images for the model. After image enhancement, a database of rare birds hitting the power line images is constructed. Second, an RBOD rare bird target detection model is constructed, and the specific task of rare bird recognition is improved through three modules: First, for the feature extraction network, the SRFD module is first used for shallow feature extraction, which retains the original feature information and fuses local information, effectively reducing the loss of feature information in the initial stage. Subsequently, the DRFD module is used as the deep feature extraction, enabling the model to capture feature information at different scales through multi-scale fusion and reducing the model's computational complexity. The use of the SRFD module and the DRFD module in the model can effectively improve the model's ability to extract features from bird images. Second, the WDBB module is used to improve the feature extraction network and the feature fusion network. With its multi-branch structure, the WDBB module enables the model to learn diverse feature representations and enhance them, effectively enhancing the model's detection ability for rare birds. Third, the MultiSEAM module is used to improve the detection module. The MultiSEAM module can accurately extract bird features in complex scenarios, enhancing the feature representation ability of the detection module. After the model is constructed, the training set and validation set images are used to train the model. Finally, the trained model is used to detect the test set images. The results show that the RBOD rare bird target detection model proposed by the present invention can accurately and quickly identify rare birds appearing around the line, providing technical support for grid operation and maintenance personnel in bird recognition and protection. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a flowchart of the method of the present invention.

[0018] Figure 2 It is a flowchart of the lightweight image super-resolution reconstruction model.

[0019] Figure 3 It is a schematic diagram of the RBOD rare bird target detection model.

[0020] Figure 4 It is a schematic diagram of the SRFD module.

[0021] Figure 5 It is a schematic diagram of the DRFD module.

[0022] Figure 6 It is a schematic diagram of the CWDB module.

[0023] Figure 7 It is a schematic diagram of the CBS module.

[0024] Figure 8 Schematic diagram of the DWBB module

[0025] Figure 9 Schematic diagram of the PSA module

[0026] Figure 10 Schematic diagram of the C2FCIB module

[0027] Figure 11 Schematic diagram of the CIB module

[0028] Figure 12 Schematic diagram of the detection module

[0029] Figure 13 Schematic diagram of the MultiSEAM module

[0030] Figure 14 Schematic diagram of the CSM module

[0031] Figure 15 Schematic diagram of the residual branch Detailed implementation manners

[0032] The following further describes the present invention in conjunction with embodiments. It is necessary to point out here that the following embodiments are only used to further illustrate the present invention and cannot be construed as limiting the protection scope of the present invention. Some non-essential improvements and adjustments made by those skilled in the art according to the above-mentioned inventive content still fall within the protection scope of the present invention.

[0033] As Figure 1 shown, the power grid wire-striking bird recognition method based on multi-scale feature enhancement includes the following steps: S1. Construct a dataset of rare bird images that accidentally strike transmission lines: According to the historical data of rare bird wire-striking accidents in the power grid, collect rare bird images through transmission line monitoring cameras and network resources. Since the transmission line monitoring cameras have low resolution and the bird images are not clear, a lightweight image super-resolution reconstruction model is used to perform super-resolution reconstruction on the rare bird images to improve the image resolution. After completing the image reconstruction, construct a dataset of rare bird images that accidentally strike transmission lines; S2. Improve based on the YOLOv10 model to construct an RBOD rare bird target detection model; S3. Train the RBOD rare bird target detection model: First, use the labelimg picture annotation tool to annotate the dataset of rare bird images that accidentally strike transmission lines constructed in step S1, and divide the dataset of rare bird images that accidentally strike transmission lines into a training set, a validation set, and a test set. Adopting the idea of transfer learning, load the pre-trained weights that have been trained, and use the divided training set and validation set to train the RBOD rare bird target detection model; S4. Use the trained RBOD rare bird target detection model to identify birds hitting the power grid line.

[0034] In step S1 of this embodiment, eleven bird species, namely black-necked crane, saker falcon, common crane, oriental white stork, great bustard, white-naped crane, little egret, white spoonbill, common kestrel, whooper swan, and northern goshawk, are selected as the identification objects. There are two hundred images of each bird species, totaling two thousand two hundred rare bird species images as the original data set. A lightweight image super-resolution reconstruction model is used to expand the original data set. Each bird species image is expanded to three hundred, totaling three thousand three hundred rare bird species image samples, and this is used as the data set for the subsequent algorithm. The lightweight image super-resolution reconstruction model is as Figure 2 shown. First, the input image is input, then image cropping, image flipping, and feature transformation are performed. Model training is carried out for the teacher model and the student model, and knowledge distillation is performed. The trained student model is used for image super-resolution reconstruction.

[0035] In this embodiment, first, according to the scientific names corresponding to the eleven bird species, labelimg is used to annotate the bird images to generate the.txt files required for model training. After the annotation is completed, the pictures and labels are divided into a training set, a validation set, and a test set according to the ratio of 8:1:1. The pre-trained weights of the model trained based on the COCO open-source large-scale image data set are loaded. The RBOD rare bird target detection model is trained using the training set and the validation set. The number of model training rounds is set to 200, the batch training size is 64, the SGD optimizer is used as the optimizer. To prevent overfitting, the weight decay is set to 0.0005, the model confidence is set to 0.7, and the initial learning rate is set to 0.01.

[0036] Specifically, the RBOD rare bird target detection model of this embodiment includes three parts: a feature extraction network, a feature fusion network, and a detection network, as Figure 3 shown. The feature extraction network consists of ten parts in sequence: the first part is the SRFD module, and the second part to the eighth part are alternately stacked by four CWDB modules and three DRFD modules (in sequence: the first CWDB module, the first DRFD module, the second CWDB module, the second DRFD module, the third CWDB module, the third DRFD module, and the fourth CWDB module). The ninth part is the SPPF module, and the tenth part is the PSA module; the output feature F 1 (large-scale feature) of the second CWDB module, the output feature F 2 (medium-scale feature) of the third CWDB module, and the output feature F 3 (small-scale feature) of the PSA module are used as the three input features of the feature fusion network.

[0037] The feature fusion network is used to effectively integrate features of different scales. The process is: F3 After passing through the first upsampling module and combined with F 2 It is fused through the first fusion module to obtain the fused feature C 1 and input into the fifth CWDB module for processing; the output feature C of the fifth CWDB module 2 After being processed by the upsampling module, it is combined with F 1 It is fused through the second fusion module to obtain the fused feature C 3 ; the fused feature C 3 After being processed by the sixth CWDB module, the output feature C of the sixth CWDB module 4 After being processed by the fourth DRFD module, it is combined with the output feature C of the fifth CWDB module 2 It is subjected to feature fusion through the third fusion module to obtain the fused feature C 5 , the fused feature C 5 is input into the seventh CWDB module for processing, and the output feature C of the seventh CWDB module 6 After being processed by the fifth DRFD module, it is combined with F 3 It is fused through the fourth fusion module to obtain the fused feature C 7 , the fused feature C 7 is input into the C2FCIB module for processing, and the output feature C of the sixth CWDB module 4 , the output feature C of the seventh CWDB module 6 , the output feature C of the C2FCIB module 8 are used as the inputs of the three detection modules of the detection network, corresponding to detections at large, medium, and small scales respectively.

[0038] It should be noted that the numbers assigned to each module in the present invention, such as the first, second, third, etc., are only for the convenience of understanding and explaining the model structure. The structures of the same type of modules with different numbers are the same, but the model parameters may be different, which is easy for those skilled in the art to understand.

[0039] For example Figure 4As shown in the figure, the SRFD module includes three stages. In the first stage, ordinary convolution is used. The features output from the first stage enter the second stage. The upper branch of the second stage consists of grouped convolution and depthwise separable convolution, and the lower branch of the second stage is pixel cutting downsampling. After the features of the upper branch and the lower branch are concatenated and convolved, they enter the third stage. The lower branch of the third stage is pixel cutting downsampling. In the upper branch of the third stage, first grouped convolution is performed, and then depthwise separable convolution and max-pooling downsampling are respectively performed through two branches. The features output from the depthwise separable convolution and the max-pooling downsampling are concatenated and convolved with the output of the lower branch of the third stage, and finally the output features of the SRFD module are obtained. Through the feature enhancement layer and multi-scale feature fusion, the SRFD module significantly improves the shallow feature extraction ability and can effectively improve the model efficiency for low-resolution and small target scenarios.

[0040] As Figure 5 shown in the figure, the DRFD module includes an upper branch and a lower branch: the lower branch of the DRFD module is pixel cutting downsampling. In the upper branch of the DRFD module, first grouped convolution is performed, and then through two branches. After the first branch completes depthwise separable convolution, the GELU activation function is used for feature enhancement, and the second branch performs max-pooling downsampling to retain key feature information. The output features of the two branches are concatenated and convolved with the output of the lower branch, and finally the output of the DRFD module is obtained. For the refined features after preliminary processing, the DRFD module provides a smoother training gradient through the GELU function, making the model easy to converge and retain more feature information, and improving the model's expression ability for deep features.

[0041] As Figure 6 shown in the figure, the processing process of the CWDB module is as follows: the image features are input into a CBS module for convolution processing, then processed by a separation module, and then sequentially processed by several WDBB modules. Finally, the features of the separation module and several WDBB modules are fused through a fusion module, and the fused features are input into another CBS module for processing to obtain the output of the CWDB module.

[0042] As Figure 7 shown in the figure, the CBS module consists of ordinary convolution (Conv), batch normalization layer, and activation function in sequence.

[0043] As Figure 8As shown in the figure, the WDBB module divides the input features into six branches for processing. The first branch is processed by a 1×1 ordinary convolution and a batch normalization layer in sequence; the second branch is processed by a 1×1 ordinary convolution, the first batch normalization layer, a k×k ordinary convolution, and the second batch normalization layer in sequence; the third branch is processed by a 1×1 ordinary convolution, the first batch normalization layer, an average pooling layer, and the second batch normalization layer in sequence; the fourth branch is processed by a k×k ordinary convolution and a batch normalization layer in sequence; the fifth branch is processed by a 1×k ordinary convolution and a batch normalization layer in sequence; the sixth branch is processed by a k×1 ordinary convolution and a batch normalization layer in sequence; the features processed by the six branches are convolved and summed, and the output features are obtained after passing through the activation function SiLU. The WDBB module enhances feature representation, reduces the computational amount, and improves the robustness and efficiency of the model through a multi-branch structure and the fusion of convolution kernels.

[0044] As Figure 9 shown, the PSA module operates as follows: The module is divided into seven parts, namely the first ordinary convolution, the multi-head self-attention module, the first feature splicing module, the feed-forward network, the second feature splicing module, the fusion module, and the second ordinary convolution. The features output by the first ordinary convolution are respectively input to the multi-head self-attention module, the first feature splicing module, and the fusion module. The features after splicing by the first feature splicing module are respectively input to the feed-forward network and the second feature splicing module. After feature fusion by the fusion module, it is input to the second ordinary convolution for convolution processing to obtain the final output of the PSA module.

[0045] As Figure 10 shown, the processing process of the C2FCIB module is as follows: The image features are input to a CBS module for convolution processing, then processed by a separation module, and then processed by several CIB modules in sequence. Finally, the features of the separation module and several CIB modules are fused through a fusion module, and the fused features are input to another CBS module for processing to obtain the output of the C2FCIB module.

[0046] As Figure 11 shown, the CIB module is composed of three depthwise separable convolutions and two CBS modules connected in series in sequence. The connection method is the first depthwise separable convolution, the first CBS module, the second depthwise separable convolution, the second CBS module, and the third depthwise separable convolution.

[0047] As Figure 12 shown, the detection module consists of two branches. The operation mode of the first branch is as follows: After the input features are convolved by the CBS module, they are processed by the MultiSEAM module and an ordinary convolution in sequence to obtain the prediction loss; the operation mode of the second branch is as follows: After the input features are processed by two consecutive CBS modules, they are processed by the MultiSEAM module and an ordinary convolution in sequence to obtain the classification loss.

[0048] As Figure 13 shown, the operation mode of the MultiSEAM module is as follows: the original input features are respectively input into three CSM modules for processing. The output features obtained by the three CSM modules are then convolved and summed with the original input features. After the summation, the features are respectively subjected to average pooling and fully connected operations. The obtained output is then multiplied by the original input features to obtain the final output features. The MultiSEAM module enables the features to pass through a multi-channel process and then through a fully connected network, integrating the information between channels and enabling the network to completely capture the spatial information.

[0049] As Figure 14 shown, the operation mode of the CSM module is as follows: after the features are input into ordinary convolution, they sequentially pass through the activation function SiLU, the batch normalization layer, and the residual branch processing. The output of the residual branch is added to the features output by the batch normalization layer, and then passes through ordinary convolution, the activation function SiLU, and the batch normalization layer to obtain the output features; As Figure 15 shown, the residual branch is sequentially composed of a depthwise separable convolution, an activation function, and a batch normalization layer. The introduction of the residual branch in the network solves the problem of gradient disappearance during training and ensures that the shallow features can be directly transmitted to the deep layer. The network can focus on learning the residuals of the input features, thereby extracting more discriminative feature information.

[0050] In this embodiment, the RBOD rare bird target detection model that has been trained in step S3 is used to detect the images in the rare bird test set. The deep learning environment used in this embodiment is: a hardware environment with Nvidia GeForce GTX 3060, 12G of video memory, and 16G of memory, as well as a software environment of Visual Studio Code 2019, CUDA 12.1, and CuDNN 8.9.6.50. The assembly language used is the python language. According to the confidence level set in step S3, the prediction boxes below the confidence threshold are removed. The model is evaluated using the conventional object detection model evaluation criteria of mean average precision (mAP) and detection speed (FPS), and a comparison is made with the YOLOv10 model. The experimental results are as follows.

[0051]

[0052] From the above results, it can be seen that compared with the YOLOv10 model, the model of this patent has obvious improvements in both detection accuracy and detection speed. Moreover, the present invention adopts a lightweight image super-resolution reconstruction model, which can also effectively improve the utilization rate of the photos taken by the transmission line monitoring camera and greatly improve the algorithm recognition efficiency. Deploying the overall model of this patent to the actual scenario can provide technical support for the operation and maintenance personnel to identify and protect birds.

[0053] The second embodiment of the present invention provides a power grid wire-striking bird recognition system based on multi-scale feature enhancement, which includes an image acquisition device and a power grid wire-striking bird recognition device. The image acquisition device is used to acquire bird images within a certain distance range of the transmission line. The power grid wire-striking bird recognition device is built-in with an RBOD rare bird target detection model for bird recognition.

[0054] The third embodiment of the present invention provides an electronic device, including a memory and a processor. The memory stores computer-readable instructions. When the instructions are executed by the processor, the processor is enabled to implement the above-mentioned power grid wire-striking bird recognition method based on multi-scale feature enhancement.

[0055] The fourth embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the above-mentioned power grid wire-striking bird recognition method based on multi-scale feature enhancement is implemented.

[0056] The above only expresses the preferred embodiments of the present invention and does not limit the present invention in other forms. Any person skilled in the art may use the disclosed content to make changes or modifications into equivalent embodiments with equivalent changes. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention still fall within the protection scope of the technical solution of the present invention.

Claims

1. A method for identifying birds striking power lines based on multi-scale feature enhancement, characterized in that: Construct and train the RBOD rare bird target detection model; use the trained RBOD rare bird target detection model to identify birds that hit the power grid; The RBOD rare bird target detection model includes three parts: feature extraction network, feature fusion network and detection network. The feature extraction network is composed of shallow feature extraction module, the first CWDB module, the first DRFD module, the second CWDB module, the second DRFD module, the third CWDB module, the third DRFD module and the fourth CWDB module, SPPF module and PSA module. The feature fusion network is used to realize the effective integration of features of different scales. The DRFD module includes an upper branch and a lower branch: the lower branch of the DRFD module is pixel cutting and downsampling. In the upper branch of the DRFD module, group convolution is first performed, and then through two branches, the first branch completes the depth-separable convolution and uses the GELU activation function for feature enhancement, and the second branch performs maximum pooling downsampling; the output features of the two branches are spliced ​​and convolved with the output of the lower branch, and finally the output of the DRFD module is obtained; The CWDB module is composed of a CBS module, a separation module, several WDBB modules, a fusion module and another CBS module in sequence.

2. The method for identifying birds striking power lines according to claim 1, characterized in that: The processing process of the feature fusion network is as follows: the output feature F3 of the PSA module is fused with the output feature F2 of the third CWDB module through the first fusion module after passing through the first upsampling module, and the obtained fusion feature C1 is input into the fifth CWDB module for processing; After being processed by the upsampling module, the output feature C2 of the fifth CWDB module is fused with the output feature F1 of the second CWDB module through the second fusion module to obtain the fused feature C3; the fused feature C3 is processed by the sixth CWDB module, and the output feature C4 of the sixth CWDB module is processed by the fourth DRFD module, and then fused with the output feature C2 of the fifth CWDB module through the third fusion module to obtain the fused feature C5, which is input to the seventh CWDB module for processing, and the output feature C6 of the seventh CWDB module is processed by the fifth DRFD module, and then fused with F3 through the fourth fusion module to obtain the fused feature C7, which is input to the C2FCIB module for processing, and the output feature C4 of the sixth CWDB module, the output feature C6 of the seventh CWDB module, and the output feature C8 of the C2FCIB module are used as the inputs of the three detection modules of the detection network.

3. The method for identifying birds striking power lines according to claim 1, characterized in that: The WDBB module divides the input features into six branches for processing. The first branch is processed by 1×1 ordinary convolution and batch normalization layer in sequence; the second branch is processed by 1×1 ordinary convolution, the first batch normalization layer, k×k ordinary convolution, and the second batch normalization layer in sequence; the third branch is processed by 1×1 ordinary convolution, the first batch normalization layer, the average pooling layer, and the second batch normalization layer in sequence; the fourth branch is processed by k×k ordinary convolution and batch normalization layer in sequence; the fifth branch is processed by 1×k ordinary convolution and batch normalization layer in sequence; the sixth branch is processed by k×1 ordinary convolution and batch normalization layer in sequence; the features processed by the six branches are convolved and summed, and the features are output after the activation function SiLU.

4. The method for identifying birds striking power lines according to claim 1, characterized in that: The operation mode of the PSA module is as follows: the module is divided into seven parts, namely the first ordinary convolution, the multi-head self-attention module, the first feature splicing module, the feedforward network, the second feature splicing module, the fusion module and the second ordinary convolution. The output features of the first ordinary convolution are respectively input into the multi-head self-attention module, the first feature splicing module and the fusion module. The features after splicing by the first feature splicing module are respectively input into the feedforward network and the second feature splicing module. After feature fusion by the fusion module, they are input into the second ordinary convolution for convolution processing to obtain the final PSA module output.

5. The method for identifying birds striking power lines according to claim 2, characterized in that: The processing process of the C2FCIB module is as follows: the image features are input into a CBS module for convolution processing, then processed by the separation module, and then processed by several CIB modules in sequence. Finally, the features of the separation module and several CIB modules are fused through the fusion module, and the fused features are input into another CBS module for processing to obtain the output of the C2FCIB module.

6. The method for identifying birds striking power lines according to claim 1, characterized in that: The shallow feature extraction module is an SRFD module, which includes three stages. The first stage adopts ordinary convolution; the output features of the first stage enter the second stage, the upper branch of the second stage is composed of grouped convolution and depth-separable convolution, the lower branch of the second stage is pixel cutting and downsampling, and the features of the upper branch and the lower branch are spliced ​​and convolved before entering the third stage; in the upper branch of the third stage, grouped convolution is first performed, and then depth-separable convolution and maximum pooling downsampling are performed respectively through two branches, the features output by the depth-separable convolution and the maximum pooling downsampling are spliced ​​and convolved with the output of the lower branch of the third stage, and finally the output features of the SRFD module are obtained.

7. The method for identifying birds striking power lines according to claim 1, characterized in that: The detection module consists of two branches. The first branch operates as follows: after the input features are convolved by the CBS module, they are processed by the MultiSEAM module and ordinary convolution in sequence to obtain the prediction loss; the second branch operates as follows: after the input features are processed by two consecutive CBS modules, they are processed by the MultiSEAM module and ordinary convolution in sequence to obtain the classification loss.

8. The method for identifying birds striking power lines according to claim 1, characterized in that: According to the historical data of rare bird collision accidents in the power grid, rare bird images are collected through transmission line monitoring cameras and network resources. A lightweight image super-resolution reconstruction model is used to reconstruct the rare bird images with super-resolution to improve the image resolution. After image reconstruction, a dataset of rare bird images that accidentally collide with transmission lines is constructed. The dataset of rare bird images that accidentally collide with transmission lines is used to train the RBOD rare bird target detection model.

9. A power grid-strike bird identification system based on multi-scale feature enhancement, characterized in that: It includes an image acquisition device and a power line-hitting bird identification device. The image acquisition device is used to collect bird images within a certain distance range of the transmission line. The power line-hitting bird identification device has a built-in RBOD rare bird target detection model for bird identification; the RBOD rare bird target detection model includes three parts: a feature extraction network, a feature fusion network and a detection network. The feature extraction network is composed of a shallow feature extraction module, a first CWDB module, a first DRFD module, a second CWDB module, a second DRFD module, a third CWDB module, a third DRFD module and a fourth CWDB module, an SPPF module and a PSA module in sequence; the feature fusion network is used to realize the effective integration of features of different scales; The DRFD module includes an upper branch and a lower branch: the lower branch of the DRFD module is pixel cutting and downsampling. In the upper branch of the DRFD module, group convolution is first performed, and then through two branches, the first branch completes the depth-separable convolution and uses the GELU activation function for feature enhancement, and the second branch performs maximum pooling downsampling; the output features of the two branches are spliced ​​and convolved with the output of the lower branch, and finally the output of the DRFD module is obtained; The CWDB module is composed of a CBS module, a separation module, several WDBB modules, a fusion module and another CBS module in sequence.

10. The power grid-strike bird identification system according to claim 9, characterized in that: The processing process of the feature fusion network is as follows: the output feature F3 of the PSA module is fused with the output feature F2 of the third CWDB module through the first fusion module after passing through the first upsampling module, and the obtained fusion feature C1 is input into the fifth CWDB module for processing; After being processed by the upsampling module, the output feature C2 of the fifth CWDB module is fused with the output feature F1 of the second CWDB module through the second fusion module to obtain the fused feature C3; the fused feature C3 is processed by the sixth CWDB module, and the output feature C4 of the sixth CWDB module is processed by the fourth DRFD module, and then fused with the output feature C2 of the fifth CWDB module through the third fusion module to obtain the fused feature C5, which is input to the seventh CWDB module for processing, and the output feature C6 of the seventh CWDB module is processed by the fifth DRFD module, and then fused with F3 through the fourth fusion module to obtain the fused feature C7, which is input to the C2FCIB module for processing, and the output feature C4 of the sixth CWDB module, the output feature C6 of the seventh CWDB module, and the output feature C8 of the C2FCIB module are used as the inputs of the three detection modules of the detection network.

Images