Corrosion diagnosis method and system for edge power transmission tower based on corrosion significance perception

Abstract

This invention belongs to the field of image data processing technology and provides a method and system for diagnosing corrosion of edge-end transmission towers based on corrosion saliency perception. The technical solution involves jointly training a prunable backbone network and a pruning controller based on acquired images of transmission tower corrosion and corresponding corrosion saliency heatmaps, resulting in a trained prunable backbone network and pruning controller. The trained prunable backbone network and pruning controller are then deployed to edge devices. Inference is performed based on real-time acquired images of the transmission tower and the deployed prunable backbone network and pruning controller to obtain the corrosion diagnosis result. The inference process includes: extracting saliency heatmap features from the real-time acquired images of the transmission tower; outputting the optimal pruning decision based on the saliency heatmap features and the constructed pruning controller; dynamically updating the channels of the backbone network based on the optimal pruning decision; and outputting the final diagnosis result based on the updated backbone network. Computational resources can be adjusted according to the complexity of the input image.

Description

Technical Field

[0001] This invention belongs to the field of image data processing technology, and in particular relates to a method and system for diagnosing corrosion of edge-end transmission towers based on the perception of corrosion salience. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] As a critical infrastructure of power transmission networks, the structural health of transmission towers directly affects the safe and stable operation of the power grid. Steel transmission towers, exposed to harsh natural environments for extended periods, are highly susceptible to corrosion. Corrosion leads to a reduction in the effective cross-section of components and a decrease in load-bearing capacity, ultimately causing structural failure and even tower collapse. Therefore, efficient and accurate corrosion detection of transmission towers has significant economic and safety implications.

[0004] In recent years, with the rapid development of computer vision and deep learning technologies, image recognition methods based on convolutional neural networks (CNNs) have gradually replaced traditional manual inspection and image processing techniques, becoming the mainstream solution for automatic diagnosis of transmission tower corrosion. These methods, by training deep models, can automatically learn the abstract features of corrosion from images, demonstrating excellent accuracy and robustness.

[0005] However, deploying advanced CNN models on edge devices (such as Jetson Xavier and Nano) with severely limited computing resources, storage space, and power consumption to achieve real-time online diagnostics still faces significant challenges. Currently, the technical approaches adopted to address this issue mainly focus on the following two aspects: 1. Employ lightweight pre-designed network models: Directly use lightweight network architectures such as MobileNet, ShuffleNet, and SqueezeNet as the backbone. These models reduce the number of parameters and computational complexity (FLOPs) to some extent through designs such as depthwise separable convolutions and channel shuffling, enabling them to run on edge devices with minimal effort.

[0006] 2. Utilize model compression techniques: After training large models (such as ResNet and VGG), static model compression techniques, such as weight quantization, knowledge distillation, and static model pruning, are employed. These techniques can compress model size and accelerate inference to a certain extent, but their compression strategies are fixed and one-size-fits-all. Once compression is complete, the model's structure and computational cost remain unchanged after deployment.

[0007] The above solutions share a common drawback: they all employ a "fixed" model structure to process all input samples. This static processing approach ignores the vast differences in the inherent complexity of input images in real-world application scenarios, leading to inefficient allocation of computational resources and bottlenecks in real-time performance, including: Resource waste for simple samples: When faced with images with simple backgrounds, no corrosion, or significant and singular corrosion areas, the fixed lightweight model still performs the entire computation process, resulting in a large amount of redundant computation. This causes inference latency and power consumption waste, failing to meet the real-time processing requirements of extremely high frame rates.

[0008] Insufficient performance for complex samples: When faced with complex images with cluttered backgrounds, weak corrosion features, scattered distribution, or poor lighting conditions, fixed lightweight models are limited by their model capacity and representation capabilities, making it difficult to perform accurate feature extraction and recognition, resulting in a significant decrease in diagnostic accuracy.

[0009] Lack of adaptive capabilities: Existing static compression or lightweight designs are offline optimizations that cannot dynamically adjust computing resources based on the real-time content of the input image, lacking intelligent, on-demand computing capabilities. Summary of the Invention

[0010] To address at least one of the technical problems mentioned above, this invention provides a corrosion diagnosis method and system for edge-end transmission towers based on corrosion saliency perception. This method considers the differences in the inherent complexity of image processing scenarios in actual applications, dynamically adjusts computing resources, and can allocate data processing as needed.

[0011] To achieve the above objectives, the present invention adopts the following technical solution: A first aspect of the present invention provides a method for diagnosing corrosion of edge-end transmission towers based on corrosion saliency perception, comprising the following steps: Acquire images of corrosion on transmission towers and generate a heat map of corrosion saliency for each image; Based on the acquired images of transmission tower corrosion and the corresponding corrosion saliency heatmaps, the constructed pruning backbone network and pruning controller are jointly trained to obtain the trained pruning backbone network and pruning controller. The trained pruning backbone network and pruning controller are deployed to edge devices. Based on real-time acquired images of the transmission tower and the deployed pruning backbone network and pruning controller, inference is performed to obtain the corrosion diagnosis results of the transmission tower. The inference process includes: extracting salient heat map features from the real-time acquired images of the transmission tower; outputting the optimal pruning decision based on the salient heat map features and the constructed pruning controller; dynamically updating the channels of the backbone network based on the optimal pruning decision; and outputting the final diagnosis result based on the updated backbone network.

[0012] Furthermore, the joint training of the constructed pruning backbone network and pruning controller based on the acquired transmission tower corrosion images and corresponding corrosion saliency heatmaps specifically includes: Statistical features are extracted from the acquired images of corrosion on the transmission towers and the corresponding corrosion saliency heatmaps, and a state vector is constructed based on the statistical features. The constructed pruning backbone network is subjected to sparse training to obtain the sparsely trained pruning backbone network. Using the constructed state vector as the state, and based on a pre-trained and sparse pruning backbone network M, a reinforcement learning paradigm is used to train the pruning controller according to the set reward function.

[0013] Furthermore, the extracted statistical features are used to construct a state vector, including the proportion of salient regions, the average strength of salientity, the salient distribution entropy, and the proportion of the largest connected regions.

[0014] Furthermore, the pruning backbone network consists of a feature extraction network and a detection head. The pruning backbone network uses ResNet-18 as the basic skeleton, and improves ResNet-18 to adapt to channel pruning. After each residual block of ResNet-18, a channel pruning gate is explicitly inserted, which directly associates the γ coefficient of each BN layer with the output channel of the previous convolutional layer as the importance score of that channel. The detection head receives multi-scale feature maps from the feature extraction network and outputs bounding box coordinates, confidence scores, and corrosion categories.

[0015] Furthermore, the constructed pruning backbone network is subjected to sparse training, including: An L1 regularization term for the γ coefficients of all BN layers was added to the loss function; The global distribution of γ values ​​in all BN layers is statistically analyzed. A pruning threshold is set, and all channels with γ values ​​below the pruning threshold are identified as redundant channels. These channels and their corresponding convolutional kernels in the next layer are then physically removed from the network.

[0016] Furthermore, the reward function is defined as follows: , in, As a precision bonus, the pruned model's detection accuracy on the current image is... β is the set weight for the efficiency reward.

[0017] Furthermore, the pruning controller is trained using a reinforcement learning paradigm based on the set reward function, including: extracting state vectors based on the sparsely trained backbone network; the controller makes decisions based on the state and outputs the pruning ratio; performing temporary pruning on the backbone network based on the decisions; and obtaining the final pruning strategy by maximizing the expected reward.

[0018] A second aspect of the present invention provides a corrosion diagnosis system for edge-end transmission towers based on corrosion saliency perception, comprising: The image acquisition module is used to acquire images of corrosion on the transmission tower and generate a heat map of corrosion salience for each image. The offline training module is used to jointly train the constructed pruning backbone network and pruning controller based on the acquired images of transmission tower corrosion and the corresponding corrosion saliency heat map, so as to obtain the trained pruning backbone network and pruning controller. The online inference module is used to deploy the trained pruning backbone network and pruning controller to edge devices. It performs inference based on real-time acquired images of the transmission tower and the deployed pruning backbone network and pruning controller to obtain the transmission tower corrosion diagnosis results. The inference process includes: extracting salient heat map features from the real-time acquired images of the transmission tower; outputting the optimal pruning decision based on the salient heat map features and the constructed pruning controller; dynamically updating the channels of the backbone network based on the optimal pruning decision; and outputting the final diagnosis result based on the updated backbone network.

[0019] A third aspect of the present invention provides a computer-readable storage medium.

[0020] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the edge-end transmission tower corrosion diagnosis method based on corrosion saliency perception as described above.

[0021] A fourth aspect of the present invention provides a computer device.

[0022] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps in the edge-end transmission tower corrosion diagnosis method based on corrosion saliency perception as described above.

[0023] Compared with the prior art, the beneficial effects of the present invention are: This invention combines the features of transmission tower corrosion images and corrosion saliency heatmaps to jointly train a constructed pruning backbone network and a pruning controller. A dynamic adjustment strategy is used to continuously adjust the pruning backbone network, taking into account the huge differences in the inherent complexity of input images in actual application scenarios. By dynamically adjusting computing resources, data processing can be allocated on demand, improving the efficient utilization of computing resources and avoiding a large amount of redundant calculations.

[0024] Advantages of additional aspects of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0025] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0026] Figure 1 This is a flowchart of the corrosion diagnosis method for edge-end transmission towers based on corrosion saliency perception provided in an embodiment of the present invention. Detailed Implementation

[0027] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0028] It should be noted that the following detailed description is illustrative and intended to provide further explanation of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0029] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0030] Example 1 like Figure 1 As shown in the figure, this embodiment provides a corrosion diagnosis method for edge-end transmission towers based on corrosion saliency perception, including the following steps: Step 1: Obtain images of corrosion on the transmission towers and generate a heat map showing the salience of corrosion for each image; In this embodiment, images of corrosion on transmission towers are acquired and finely annotated (e.g., pixel-level segmentation labels or bounding box labels). A pre-trained saliency detection model (e.g., U-Net) or traditional algorithms (e.g., FT, HC) is used to generate a corresponding saliency map for each image in the training set. This heatmap has the same size as the original image, and each pixel value represents the saliency (0~1) of corrosion at that location.

[0031] Step 2: Based on the acquired images of transmission tower corrosion and the corresponding corrosion saliency heatmap, the constructed pruning backbone network and pruning controller are jointly trained to obtain the trained pruning backbone network and pruning controller. Specifically, the steps include the following: Step 201: Extract statistical features based on the acquired images of corrosion on the transmission tower and the corresponding heat map of corrosion significance, and construct a state vector based on the statistical features; In this embodiment, in the corrosion saliency heatmap, each pixel value represents the degree of corrosion presence at that location. Instead of directly using a high-dimensional heatmap, compact and information-rich statistical features are extracted from it as a state vector s describing the "diagnostic difficulty" of the image. This state vector will serve as the input to the pruning controller.

[0032] Specifically, the extracted statistical features include the proportion of significant regions, the average intensity of significance, the significance distribution entropy, and the proportion of the largest connected regions; Among them, the proportion of significant areas The calculation formula is: , in, The pixels representing the rusted areas in the rust saliency heatmap. The set binarization threshold, This represents the total number of pixels; the higher the percentage of significant areas, the larger and more obvious the rusted area, and the more "simple" the image may be, tending towards pruning.

[0033] Among them, the average significance intensity The calculation formula is: , in, It is greater than in heatmap M In the part where the average significance level is higher, it indicates that the model has a higher confidence level in the rusted area, and the image may be more "simple".

[0034] Among them, the saliency distribution entropy Calculate the normalized histogram p of the heatmap M, and then calculate the entropy: , in, The first in the histogram i The probability value of an interval is determined by its entropy. A high entropy value indicates that the salience is scattered in many places, and the background may be cluttered; a low entropy value indicates that the salience is concentrated, and the image is "cleaner". Low-entropy images tend to be pruned.

[0035] The percentage of the largest connected region is: On the binarized (M>threshold_T) heatmap, find the largest connected region and calculate the ratio of its area to the image area. A high value indicates the presence of a dominant, continuous rust region; a low value indicates that the rust is scattered and more difficult to diagnose. High values ​​tend to indicate pruning.

[0036] Finally, the constructed state vector is represented as This state vector effectively encodes the complexity of the current image.

[0037] Step 202: Perform sparse training on the constructed pruning backbone network to obtain the sparsely trained pruning backbone network M; In this embodiment, the pruning backbone network aims to balance high accuracy and low latency, and its core consists of a feature extraction network and a detection head.

[0038] The prunable backbone network uses ResNet-18 as its basic skeleton, but improves upon it to accommodate channel pruning. An explicit channel pruning gate is inserted after each residual block of ResNet-18. This gate is not an additional complex module, but rather implemented by utilizing the scaling factor (γ parameter) inherent in the batch normalization (BN) layers. Specifically, the γ coefficients of each BN layer are directly correlated with the output channels of the previous convolutional layers as an importance score for that channel. These γ coefficients are subject to sparsity constraints during training, allowing the network to automatically learn the importance and redundancy of channels. To locate and classify the rusted areas, a lightweight YOLO detection head, such as one based on YOLOv3 or YOLOX, is used. This detection head receives multi-scale feature maps from a feature extraction network and outputs bounding box coordinates, confidence scores, and rust categories.

[0039] As a further implementation, the constructed pruning backbone network is subjected to sparse training, including: Specifically, a structured pruning method based on the BN layer scaling factor is adopted, and the specific steps are as follows: Sparsity training: During model training, in addition to the regular classification and regression losses, an L1 regularization term for the γ coefficients of all BN layers is added to the loss function. This forces the network to push the γ values ​​of unimportant channels towards zero during optimization.

[0040] This embodiment uses a norm criterion, such as L1-Norm, which calculates the sum of the absolute values ​​of all weights of each convolutional kernel (corresponding to one output channel) in a convolutional layer. The smaller the norm (size) of a channel's weights, the closer its activation value is to zero, and the smaller its contribution to the next layer. Therefore, it is less important, and unimportant channels can be removed by setting a threshold.

[0041] Channel pruning: The global distribution of γ values ​​across all BN layers is statistically analyzed; a pruning threshold is set (e.g., after sorting all γ values, the kth percentile value is used as the threshold), and all channels with γ values ​​below this threshold are identified as redundant channels. Subsequently, these channels and their corresponding convolutional kernels in the next layer are physically removed from the network, forming a smaller, more compact "sub-network".

[0042] Pruned backbone network fine-tuning: Pruning temporarily reduces the model's accuracy. To restore its performance, the pruned, lightweight model is fine-tuned for one epoch using the training data. During this stage, the L1 regularization constraint is removed, and only the detection loss is used, allowing the model weights to adapt to the new compact structure.

[0043] The loss function formula is: , where λ is the sparsity penalty coefficient, used to control the intensity of pruning.

[0044] Step 203: Using the constructed state vector as the state, and based on the pre-trained and sparse pruning backbone network M, the pruning controller is trained using the reinforcement learning paradigm based on the set reward function. In this embodiment, a reinforcement learning paradigm is used to train the pruning controller. Specifically, in the environment of the pre-trained and sparse pruning backbone network M, the feature vector s extracted from the saliency heatmap is used as the state, and the specific action is the pruning ratio a. The defined reward function specifically includes an accuracy reward R_accuracy and an efficiency reward R_efficiency, expressed as: , in, For accuracy rewards, the detection accuracy of the pruned model on the current image can be measured by the confidence of the detection head or IoU, for example, by using the difference or relative value of mAP; To incentivize higher pruning rates, an efficiency reward can be set. This can be calculated using R_efficiency = a (the action itself) or (1 - FLOPs_after_pruning / FLOPs_original), where β is a weighting coefficient that can be preset.

[0045] The specific training process employs policy gradient algorithms such as proximal policy optimization or REINFORCE. State vectors are extracted from the already sparsely trained backbone network. The controller makes decisions based on the state, outputs the pruning ratio, and performs temporary pruning on the backbone network according to the decisions. The final pruning policy is obtained by maximizing the expected reward.

[0046] Step 3: Deploy the trained pruning backbone network and pruning controller to the edge device, and perform inference based on the real-time acquired images of the transmission tower and the deployed pruning backbone network and pruning controller to obtain the corrosion diagnosis results of the transmission tower. This invention constructs a collaborative recognition model of "coarse localization by the backend server and fine judgment by the edge device." The backend server is responsible for executing the computationally intensive process of generating a saliency heatmap of corrosion and sparse training, outputting the trained model and pruning strategy. The edge device deploys the resulting lightweight, prunable backbone network. During inference, the device dynamically activates important channels retained after pruning based on the features of the input image, skipping redundant channels, thereby achieving adaptive and rapid inference.

[0047] In this embodiment, the reasoning based on real-time acquired images of the transmission towers and the deployed pruning backbone network and pruning controller includes: Based on real-time acquired images of transmission towers, salient heatmap features are extracted. Based on these features and the constructed pruning controller, the optimal pruning decision is output. The corresponding important channels in the backbone network are dynamically activated and unimportant channels are blocked according to the optimal pruning decision. Finally, the final diagnostic result is output through the backbone network.

[0048] Example 2 This embodiment provides a corrosion diagnosis system for edge-end transmission towers based on corrosion saliency perception, including: The image acquisition module is used to acquire images of corrosion on the transmission tower and generate a heat map of corrosion salience for each image. The offline training module is used to jointly train the constructed pruning backbone network and pruning controller based on the acquired images of transmission tower corrosion and the corresponding corrosion saliency heat map, so as to obtain the trained pruning backbone network and pruning controller. The online inference module is used to deploy the trained pruning backbone network and pruning controller to edge devices. It performs inference based on real-time acquired images of the transmission tower and the deployed pruning backbone network and pruning controller to obtain the transmission tower corrosion diagnosis results. The inference process includes: extracting salient heat map features from the real-time acquired images of the transmission tower; outputting the optimal pruning decision based on the salient heat map features and the constructed pruning controller; dynamically updating the channels of the backbone network based on the optimal pruning decision; and outputting the final diagnosis result based on the updated backbone network.

[0049] It should be noted that the specific implementation of the corrosion diagnosis system for edge-end transmission towers based on corrosion saliency perception in this embodiment of the invention is similar to the specific implementation of the corrosion diagnosis method for edge-end transmission towers based on corrosion saliency perception in this embodiment of the invention. For details, please refer to the description in the method section. To reduce redundancy, it will not be repeated here.

[0050] Example 3 This embodiment provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps in the edge-end transmission tower corrosion diagnosis method based on corrosion saliency perception as described above.

[0051] Example 4 This embodiment provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps in the edge-end transmission tower corrosion diagnosis method based on corrosion saliency perception as described above.

[0052] Example 5 This embodiment provides a program product, which is a computer program product, including a computer program. When the computer program is executed by a processor, it implements the steps in the edge-end transmission tower corrosion diagnosis method based on corrosion saliency perception as described above.

[0053] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of hardware embodiments, software embodiments, or embodiments combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.

[0054] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0055] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0056] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0057] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0058] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A corrosion diagnosis method for edge-end transmission towers based on corrosion saliency perception, characterized in that, Includes the following steps: Acquire images of corrosion on transmission towers and generate a heat map of corrosion saliency for each image; Based on the acquired images of transmission tower corrosion and the corresponding corrosion saliency heatmaps, the constructed pruning backbone network and pruning controller are jointly trained to obtain the trained pruning backbone network and pruning controller. The trained pruning backbone network and pruning controller are deployed to edge devices. Based on real-time acquired images of the transmission tower and the deployed pruning backbone network and pruning controller, inference is performed to obtain the corrosion diagnosis results of the transmission tower. The inference process includes: extracting significant heat map features from the real-time acquired images of the transmission tower; outputting the optimal pruning decision based on the significant heat map features and the constructed pruning controller; dynamically updating the channels of the backbone network based on the optimal pruning decision; and outputting the final diagnosis result based on the updated backbone network. The method involves jointly training the constructed pruning backbone network and pruning controller based on the acquired images of transmission tower corrosion and the corresponding corrosion saliency heatmaps, specifically including: Statistical features are extracted from the acquired images of corrosion on the transmission towers and the corresponding corrosion saliency heatmaps, and a state vector is constructed based on the statistical features. The constructed pruning backbone network is subjected to sparse training to obtain the sparsely trained pruning backbone network. Using the constructed state vector as the state, and based on the pre-trained and sparse pruning backbone network M, the pruning controller is trained using a reinforcement learning paradigm based on the set reward function. The extracted statistical features are used to construct a state vector, which includes the proportion of salient regions, the average strength of salientity, the salient distribution entropy, and the proportion of the largest connected regions.

2. The corrosion diagnosis method for edge-end transmission towers based on corrosion saliency perception as described in claim 1, characterized in that, The pruning backbone network consists of a feature extraction network and a detection head. The pruning backbone network uses ResNet-18 as the basic skeleton and improves ResNet-18 to adapt to channel pruning. After each residual block of ResNet-18, a channel pruning gate is explicitly inserted, which directly associates the γ coefficient of each BN layer with the output channel of the previous convolutional layer as the importance score of that channel. The detection head receives multi-scale feature maps from the feature extraction network and outputs bounding box coordinates, confidence scores, and corrosion categories.

3. The corrosion diagnosis method for edge-end transmission towers based on corrosion saliency perception as described in claim 1, characterized in that, The constructed pruning backbone network is then subjected to sparse training, including: An L1 regularization term for the γ coefficients of all BN layers was added to the loss function; The global distribution of γ values ​​in all BN layers is statistically analyzed. A pruning threshold is set, and all channels with γ values ​​below the pruning threshold are identified as redundant channels. These channels and their corresponding convolutional kernels in the next layer are then physically removed from the network.

4. The corrosion diagnosis method for edge-end transmission towers based on corrosion saliency perception as described in claim 1, characterized in that, The reward function is defined as follows: ， in, As a precision bonus, the pruned model's detection accuracy on the current image is... β is the set weight for the efficiency reward.

5. The corrosion diagnosis method for edge-end transmission towers based on corrosion saliency perception as described in claim 1, characterized in that, The pruning controller is trained using a reinforcement learning paradigm based on a set reward function. This includes: extracting state vectors from a sparsely trained backbone network; the controller making decisions based on the states and outputting a pruning ratio; temporarily pruning the backbone network based on the decisions; and obtaining the final pruning strategy by maximizing the expected reward.

6. A corrosion diagnosis system for edge-end transmission towers based on corrosion saliency perception, characterized in that, include: The image acquisition module is used to acquire images of corrosion on the transmission tower and generate a heat map of corrosion salience for each image. The offline training module is used to jointly train the constructed pruning backbone network and pruning controller based on the acquired images of transmission tower corrosion and the corresponding corrosion saliency heat map, so as to obtain the trained pruning backbone network and pruning controller. The online inference module is used to deploy the trained pruning backbone network and pruning controller to edge devices. It performs inference based on real-time acquired images of the transmission tower and the deployed pruning backbone network and pruning controller to obtain the corrosion diagnosis results of the transmission tower. The inference process includes: extracting significant heat map features from the real-time acquired images of the transmission tower; outputting the optimal pruning decision based on the significant heat map features and the constructed pruning controller; dynamically updating the channels of the backbone network based on the optimal pruning decision; and outputting the final diagnosis result based on the updated backbone network. The method involves jointly training the constructed pruning backbone network and pruning controller based on the acquired images of transmission tower corrosion and the corresponding corrosion saliency heatmaps, specifically including: Statistical features are extracted from the acquired images of corrosion on the transmission towers and the corresponding corrosion saliency heatmaps, and a state vector is constructed based on the statistical features. The constructed pruning backbone network is subjected to sparse training to obtain the sparsely trained pruning backbone network. Using the constructed state vector as the state, and based on the pre-trained and sparse pruning backbone network M, the pruning controller is trained using a reinforcement learning paradigm based on the set reward function. The extracted statistical features are used to construct a state vector, which includes the proportion of salient regions, the average strength of salientity, the salient distribution entropy, and the proportion of the largest connected regions.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps in the corrosion diagnosis method for edge-end transmission towers based on corrosion saliency perception as described in any one of claims 1-5.

8. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps in the edge-end transmission tower corrosion diagnosis method based on corrosion saliency perception as described in any one of claims 1-5.

Images