A power tower component detection method for unmanned airport scenarios

Abstract

The application discloses a kind of unmanned airport scene-oriented power pole tower component detection method.The method is by constructing the target detection model comprising adaptive multi-receptive field module, hierarchical wavelet interaction unit and distributed perception confidence optimization head, solve the key technical problems in unmanned aerial vehicle power inspection.In the feature extraction stage, parallel multi-dilation deep convolution and channel attention mechanism are used to enhance the multi-scale feature perception ability;In the feature fusion stage, through wavelet transform decomposition and cross-scale channel interaction, semantic alignment and edge detail preservation are realized;In the prediction stage, based on the statistical analysis of bounding box distribution, the classification confidence is adaptively calibrated, and the consistency of positioning accuracy and classification score is improved.The application effectively overcomes the problem of missed detection and false detection caused by large scale difference of components and complex background, significantly improves the accuracy and reliability of the detection result, and provides effective technical support for unmanned aerial vehicle power inspection.

Description

Technical Field

[0001] This invention relates to a method for detecting power pole components in drone scenarios, belonging to the field of computer vision. Background Technology

[0002] As the core energy source for modern society, the stable supply of electricity directly depends on the safe operation of high-voltage transmission lines. As critical infrastructure of these lines, power poles are constantly exposed to complex natural environments, requiring regular inspection and maintenance to ensure grid reliability. Traditional manual inspection methods suffer from inherent drawbacks such as high labor intensity, low inspection efficiency, and high safety risks, especially in complex terrains like mountainous and forested areas where manual inspections are even more challenging. In recent years, with the rapid development of drone technology, drone-based power inspection solutions have demonstrated significant advantages. Drones, with their flexible mobility, wide monitoring perspective, and efficient operating modes, provide a completely new technological means for power inspection. This inspection method not only greatly improves operational efficiency but also effectively reduces personnel safety risks, becoming an important development direction for power facility inspection.

[0003] In UAV power line inspection missions, accurately identifying various key components on power poles, including insulators, fittings, and bolts, is a fundamental step in assessing equipment condition. Compared to overall pole identification, component-level inspection faces more severe technical challenges: First, components vary significantly in size, ranging from several-meter-long insulator strings to centimeter-sized fasteners coexisting in the same scene, placing extremely high demands on the multi-scale feature perception capabilities of the inspection model; second, components have diverse shapes, making them susceptible to changes in angle, partial obstruction, and lighting conditions from an aerial perspective; furthermore, complex environmental backgrounds such as mountains, fields, and buildings, as well as weather changes, further increase the difficulty of accurate inspection. These factors collectively constitute the technical bottleneck of power component inspection.

[0004] Currently, deep learning-based object detection methods have been widely applied in the field of power component detection. Researchers have proposed various improvement strategies, including introducing attention mechanisms to enhance feature representation capabilities, using feature pyramid networks to fuse multi-scale information, and optimizing detection performance by improving network structure and loss functions. However, these methods still have significant shortcomings in several key aspects: at the feature extraction level, the receptive field design of traditional convolutional neural networks is relatively fixed, making it difficult to adapt to the feature learning needs of components at different scales. This limitation leads to the model's inability to simultaneously and effectively capture significantly different target features such as large insulators and micro bolts, resulting in a high false negative rate for small-scale components, while the feature representation of large-scale components is also insufficient. Tan et al. used directional object detection and deformable convolutional structures to achieve high-precision identification of transmission towers in remote sensing images. Li et al. proposed a cross-scale spatial attention detector, which effectively identifies minute components in the mechanical connections of transmission lines, but it does not involve adaptive adjustment of the receptive field. In terms of feature fusion, existing multi-scale fusion methods mostly use simple upsampling and feature concatenation operations, lacking in-depth mining of semantic relationships between cross-scale features. This fusion method is prone to feature mismatch problems in complex scenarios, resulting in the loss of key details during transmission, which directly affects the ability to locate the precise boundaries of components.

[0005] At the prediction output level, there is a significant mismatch between classification confidence and localization quality. Existing methods lack an accurate evaluation mechanism for the quality of predicted bounding boxes, leading to situations where correct detections with accurate localization but low confidence are filtered out, while incorrect detections with inaccurate localization but high confidence are retained, severely impacting the reliability of the detection results. Wang et al. proposed further optimizing the accuracy and efficiency of small target detection through a dual prediction head structure and coordinate attention mechanism. Furthermore, existing improvement schemes often focus on optimizing a single technical aspect, lacking a systematic solution. Most methods, while improving the performance of one aspect, may negatively impact other aspects or significantly increase computational complexity, making it difficult to achieve ideal overall results in practical inspection scenarios. These problems restrict further improvement in the efficiency of power line inspection, necessitating an innovative technical solution that can systematically optimize the entire process from feature extraction to result prediction.

[0006] In summary, current methods for detecting power pole components still have significant shortcomings in the three core stages of feature extraction, feature fusion, and detection prediction: First, the receptive field structure design in the feature extraction stage is simplistic, lacking multi-scale feature perception capabilities and failing to adapt to the vast scale differences of power pole components. Second, the feature fusion stage lacks a semantic alignment mechanism for cross-scale features, resulting in insufficient detail preservation and decreased feature discrimination in complex backgrounds. Third, the detection prediction stage fails to effectively align classification confidence with localization accuracy and lacks accurate evaluation of the predicted bounding box quality, affecting the reliability of the detection results. Therefore, there is an urgent need for a precise detection method for power pole components in UAV scenarios, which can solve the above-mentioned technical challenges through systematic structural innovation and meet the practical application needs of automated power inspection. Summary of the Invention

[0007] The purpose of this invention is to provide a method for detecting power pole components in unmanned aerial vehicle (UAV) scenarios, comprising the following steps: This invention is implemented using the following scheme: S1: Obtain a dataset of drone aerial images of power poles; S2: Preprocess the power pole and tower dataset; S3: Establish a target detection model for power pole components in a drone scenario; S3.1: In the feature extraction stage, an adaptive multi-receptive field module is adopted. This module performs multi-scale feature extraction through its internal parallel multi-dilation depth convolution sub-module. This sub-module covers receptive fields of different scales on the same channel dimension through multiple pre-set convolution branches with fixed dilation rates, so as to enhance the ability to model the context of the target. Subsequently, the feature is reorganized by combining the channel attention sub-module and the dimension transformation sub-module, and with the help of residual connections, the enhanced multi-scale features are output. S3.2: In the feature fusion stage, a hierarchical wavelet interaction unit is used to perform wavelet transform decomposition, detail enhancement, and channel interaction on features at adjacent scales. First, wavelet transform decomposition is performed on large-scale features to obtain low-frequency and high-frequency components in the horizontal, vertical, and diagonal directions. High-frequency details are enhanced by fusion, and low-frequency components are interactively compressed with small-scale features in the channel dimension. Then, the fused features are reconstructed into a unified scale through inverse wavelet transform, thereby achieving semantic alignment and preserving edge details in complex backgrounds. S3.3: In the prediction stage, a distribution-aware confidence optimization head is used. This module performs statistical analysis on the distribution of the predicted bounding boxes, adjusts the initial classification score based on the Top-k probabilities and their mean statistical features, and thus optimizes the confidence of the target. This process uses a multilayer perceptron to process the statistical features, generates the optimized confidence, and combines the bounding box coordinate distribution features to output the final detection box and class probability. S4: Train the model using the training set of the dataset to obtain the optimal model; S5: Use the optimal model to test the test set images of the dataset to obtain the final test results.

[0008] Furthermore, the InsPLAD remote sensing dataset is cropped into 640x640 image blocks without sliding window slicing or block overlap, and zero-padding is applied to areas less than 640 pixels to 640x640; the label files are converted to YOLO detection format and coordinate mapping and normalization are performed based on the scaling and padding relationship.

[0009] Furthermore, the adaptive multi-receptive field module consists of a multi-expanded depth convolution submodule, a channel attention submodule, and a dimension transformation submodule.

[0010] Furthermore, the adaptive multi-receptive field module is an improvement on the C3k2 framework, replacing the bottleneck unit with a multi-expanded depth convolution submodule. This submodule is customized for different convolution kernel sizes. When the convolution kernel size is ≥7, multi-branch expanded depth convolution is used, and its expansion rate combination is adapted according to the convolution kernel size. When the convolution kernel size is 7, a multi-branch convolution structure with an expansion rate of [1,2,3] is used. For a size of 13, an expansion rate combination of [1,2,3,4,5] is used to adapt to the detection requirements of power pole tower components of different sizes.

[0011] Furthermore, the adaptive multi-receptive field module also includes a channel attention submodule and a dimension transformation submodule. The channel attention submodule is configured to adaptively generate and weight the channel weights of the input feature map. It compresses the spatial dimension of the feature map through a global average pooling layer, generates a weight vector representing the importance of each channel using convolution operations and ReLU and Sigmoid activation functions, and multiplies the weight vector with the original input feature map channel by channel. The module also includes a dimension transformation submodule. This submodule transforms the feature tensor from batch-channel-height-width format to batch-height-width-channel format through a layout transformation from NCHW to NHWC. It expands the feature dimension using a linear layer and introduces GELU activation. It then normalizes the global response of the expanded features through a global response normalization layer to stabilize the feature distribution, and finally compresses the feature dimension through a linear layer. Finally, it restores the feature format to batch-channel-height-width format through an inverse transformation from NHWC to NCHW.

[0012] Furthermore, the hierarchical wavelet interaction unit is configured to fuse a large-scale feature and a small-scale feature. It decomposes the large-scale feature into low-frequency components and high-frequency components in three directions (horizontal, vertical, and diagonal) through wavelet transform, enhances the high-frequency components, performs channel interaction between the low-frequency components and the small-scale feature, and finally reconstructs the processed components into a unified-scale fused feature through inverse wavelet transform.

[0013] Furthermore, the wavelet transform in the hierarchical wavelet interaction unit decomposes the large-scale features into approximately low-frequency components and high-frequency components in three directions: horizontal, vertical, and diagonal. The high-frequency components in the three directions are concatenated and fused and enhanced through a residual block containing continuous 3x3 convolutions. The approximately low-frequency components and the small-scale features are concatenated in the channel dimension and interacted and compressed through a channel transform module composed of 1x1 convolutions, compressing the size to three times the dimension of the large-scale features to meet the input dimension requirements of the inverse wavelet transform. The enhanced high-frequency features are concatenated with the interactively compressed low-frequency small-scale features and fused through the inverse wavelet transform to obtain the fused features.

[0014] Furthermore, the distribution-aware confidence optimization head is configured to perform the following operations: Parallel processing of the input features; generating discrete coordinate distribution features of the bounding box through a first convolutional branch, the first convolutional branch containing two consecutive 3x3 standard convolutional layers and a final 1x1 convolutional layer; and generating an initial classification score through a second convolutional branch; reshaping the discrete coordinate distribution features into a tensor structure of [B, M, 4, H, W], where M is the regression maximum value, and processing it with a Softmax function to obtain the probability distribution of the bounding box coordinates; and extracting the coordinates of each coordinate dimension from the probability distribution. The first K maximum probability values ​​are calculated, and the mean of these K probability values ​​is calculated. The first K probability values ​​are concatenated with the mean to form a distribution statistical feature. A multilayer perceptron is used to perform a nonlinear transformation on the distribution statistical feature to generate a quality score adjustment value. The quality score adjustment value is added to the initial classification score to output the optimized classification confidence. The second convolutional branch adopts a hierarchical structure that includes depthwise separable convolutions and standard convolutions. The value of K is 4. The multilayer perceptron consists of two 1x1 convolutional layers, which are used to further perform nonlinear transformations on the statistical features. Its hidden layer dimension is 64.

[0015] The beneficial effects of this invention are: This invention proposes a novel adaptive multi-receptive field module. Through the collaborative design of customized expansion rate combinations ([1,2,3] for kernel=7, [1,2,3,4,5] for kernel=13) and a channel attention mechanism, it simultaneously captures features of large-scale structures such as power poles and towers, as well as small-scale components such as insulators and bolts. This overcomes the limitation of traditional models' difficulty in multi-scale adaptation and significantly enhances multi-scale feature extraction capabilities. Furthermore, this invention proposes a novel hierarchical wavelet interaction unit. It achieves high- and low-frequency feature decoupling through Haar wavelet transform, and, in conjunction with residual blocks, enhances high-frequency details and 1× 1. Convolutional optimization of cross-scale interaction effectively mitigates interference from complex backgrounds such as sky and vegetation in drone aerial photography, significantly improving feature discrimination in complex scenes; This invention proposes a novel distribution-aware confidence optimization head, which mines the Top-4 statistical features of the bounding box probability distribution and adaptively calibrates the initial classification score using MLP, addressing the pain point of "mismatch between confidence and positioning accuracy" for small targets and improving positioning stability; This invention validates the effectiveness and superiority of the method on the InsPLAD power pole component detection dataset, and uses comprehensive evaluation metrics to evaluate model accuracy: CDP, CDR, and Experimental results show that the present invention outperforms the YOLOv11 baseline model. Increased by 4.4%, The method achieves a 7.0% improvement and outperforms mainstream models such as YOLOv3-Tiny, YOLOv8n, and RT-DETR-L, fully demonstrating the effectiveness of the method in the power pole component inspection task. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings.

[0017] Figure 1 This is a flowchart illustrating the overall implementation of this invention. Figure 2 This is a framework diagram of the overall network model in an embodiment of the present invention; Figure 3 This is a schematic diagram of the adaptive multi-receptive field module structure in an embodiment of the present invention; Figure 4 This is a schematic diagram of the hierarchical wavelet interaction unit structure in an embodiment of the present invention; Figure 5 This is a schematic diagram of the distributed sensing confidence optimization head structure in an embodiment of the present invention. Detailed Implementation

[0018] The technical solutions in the embodiments of the present invention will be further described below with reference to the accompanying drawings, which are not intended to limit the present invention.

[0019] like Figure 1 As shown, Figure 1 As shown, an embodiment of the present invention provides a method for detecting power pole components in a drone scenario, including steps S1-S6: S1: Obtain the drone aerial image dataset of power pole towers, preprocess it, and input it into the built model; S2: The backbone network design uses an adaptive multi-receptive field module to extract multi-scale features through parallel multi-dilated convolutions; The adaptive multi-receptive field module includes a multi-expansion depth convolution submodule, a channel attention submodule, and a dimensionality transformation submodule, along with residual connections. The multi-expansion depth convolution submodule covers receptive fields of different scales through multi-branch dilated convolution, enhancing the context modeling capability for power pole components of different sizes. The channel attention submodule strengthens identity-related features and suppresses background interference by generating channel weights. The dimensionality transformation submodule stabilizes feature distribution to adapt to subsequent processing through feature format and dimensionality optimization. Finally, residual connections are combined to preserve the original feature information, alleviating the problem of deep training. Overall, it is used to adapt to multi-scale feature extraction of power pole components in UAV scenarios, outputting enhanced multi-scale features.

[0020] S3: Feature fusion network, which uses hierarchical wavelet interaction units to fuse multi-scale features, combining high-frequency and low-frequency features; The hierarchical wavelet interaction unit includes a wavelet transform decomposition module, a high-frequency enhancement module, a cross-scale interaction module, and a wavelet inverse transform reconstruction module. The wavelet transform decomposition module performs multi-directional decomposition on large-scale features, separating low-frequency structural information from high-frequency detail information. The high-frequency enhancement module strengthens target edge and texture features in high-frequency components. The cross-scale interaction module realizes channel association and adaptation between low-frequency information and small-scale features. Finally, the wavelet inverse transform reconstruction module fuses the processed high and low frequency features into unified scale features, enhancing the semantic consistency and detail preservation capability of power pole tower component features in complex backgrounds.

[0021] S4: In the prediction phase, a distributed-aware confidence optimization head is used to optimize the confidence of the detection boxes; The distributed perception confidence optimization includes a parallel feature processing branch, a distribution statistics module, a confidence adjustment module, and a detection output fusion module. The parallel feature processing branch generates bounding box coordinate distribution features and initial classification scores, respectively. The distribution statistics module performs probability analysis and feature extraction on the coordinate distribution to capture positioning quality correlation information. The confidence adjustment module calibrates the initial classification score based on the distribution statistics features. Finally, the detection output fusion module integrates the optimized confidence and bounding box information to improve the matching degree between the confidence and positioning accuracy of power pole component detection in UAV scenarios.

[0022] S5: Train the model using the training set and verify the detection accuracy using the test set.

[0023] Furthermore, in step 1, the InsPLAD remote sensing dataset is cropped into 640x640 image blocks without sliding window slicing or block overlap, and zero-padding is used to bring the area to 640x640 for areas less than 640 pixels; the label file is converted to YOLO detection format and coordinate mapping and normalization are completed according to the scaling and padding relationship.

[0024] like Figure 2 The diagram shown illustrates the overall framework of an improved target detection method for detecting power pole components for unmanned aerial vehicles (UAVs) provided by this invention. It mainly comprises three parts: a backbone network, a feature fusion section, and a detection head. The backbone network includes multiple convolutional layers, an adaptive multi-receptive field module, a C3k2 module, an SPPF module, and a C2PSA module for multi-scale feature extraction and enhancement. The feature fusion section includes a splicing layer, convolutional layers, a C3k2 module, and a hierarchical wavelet interaction unit for semantic alignment and detail fusion of cross-scale features. The detection head section features a distributed perceptual confidence optimization head, used to optimize classification confidence and output the final detection result for the power pole components.

[0025] The present invention provides a preferred embodiment for performing S2 to obtain an enhanced multi-scale feature map. For example... Figure 3 As shown, it includes the following steps: S2-1: The multi-dilation depthwise convolutional submodule extracts multi-scale basic features from the input feature map. ,in The number of hidden channels is determined by the number of output channels. With compressibility Calculated Multi-scale receptive fields are fused using an expanded depthwise convolutional structure with main branches and multiple auxiliary branches. The process can be represented as follows: The main branch is: , Auxiliary branch calculation: , , , Multi-branch fusion: , in, The kernel size for the three auxiliary branches. , This represents the expansion rate of the corresponding auxiliary branch, where , Indicates the kernel size as A depthwise convolution with an expansion rate of d; For batch normalization operation; This is the multi-scale basic feature map output after multi-branch fusion.

[0026] S2-2: The channel attention submodule enhances key features, improving the multi-scale base feature map. After batch normalization, channel weight calibration is performed to enhance the characteristics of tower components and suppress background interference. First, global average pooling compresses the spatial dimension: , For global average pooling, the spatial dimension is compressed to Focusing on channel-level information; Next use Convolution reduces the number of channels from Compressed to the number of internal neurons : ,in ,for Convolution operation; Next, a nonlinear representation is introduced through ReLU to enhance feature selectivity: ; Next, channel-expanded convolution is performed: using Convolution reduces the number of channels from Expand back Restore the original channel dimensions: ; Next, weight normalization is performed, and the channel weights are mapped to the [0,1] interval using the Sigmoid function to generate interpretable attention weights: ; Finally, feature weighting fusion is performed: the original features are multiplied element-wise by the channel weights to enhance the key channel features. , This is element-wise multiplication. This is the attention-weighted feature map.

[0027] The S2-3 dimensional transformation submodule optimizes the feature distribution and applies attention-weighted features. The feature distribution is optimized through feature format conversion, dimensionality scaling, and normalization operations. The specific process is as follows: First, the dimension order of the feature tensor is adjusted from batch, channel, height, width to batch, height, width, channel to adapt to the linear transformation of the channel dimension; then, the number of channels is changed from... Expansion to The process can be represented as: , in, For format conversion, For channel linear transformation, This is the activation function.

[0028] Then, through channel compression and inverse format conversion: the linear layer reduces the number of channels from... Compress back Finally, it is converted back to NCHW format, and after compression, it needs to undergo additional batch normalization to stabilize the training. , , , The final output is a feature map with optimized dimensions. This process enhances the global discriminative power of key features through global response normalization, and combines dimensional scaling and normalization to stabilize training, providing high-quality input for subsequent residual fusion.

[0029] S2-4 Residual Fusion Output: Features after dimensionality optimization Compared with the original input features Residual fusion is performed to preserve the original feature information and enhance multi-scale feature representation. The specific process is as follows: , Final output This process preserves the basic information of the original features through the residual structure, while integrating the optimized multi-scale features, so that the output features can be more accurately adapted to the feature extraction needs of power poles.

[0030] This invention provides a preferred embodiment for executing S3, which is used for the fusion enhancement of multi-scale feature maps. It achieves deep interaction between high-frequency details of large-scale features and fine-grained information of small-scale features through wavelet transform, such as... Figure 4 As shown, it includes the following steps: S3-1 wavelet transform module for large-scale feature multi-component separation: First, the high-rise large-scale feature map The input wavelet transform submodule is decomposed into 4 groups of semantically clear feature components through 2×2 grouped convolution with fixed weights. ; The core of the wavelet transform submodule WT consists of 4 groups. Convolution kernels, with weights as follows: low-frequency component convolution kernel Horizontal high-frequency component convolution kernel Vertical high-frequency component convolution kernel Diagonal high-frequency component convolution kernel ; Through channel separation operation It is divided into four components; ; Where the dimensions of a, h, v, and d are all This allows for a 50% reduction in space size while retaining critical information.

[0031] S3-2 high-frequency component residual enhancement adds horizontal, vertical, and diagonal high-frequency components element-wise along the channel dimension and inputs them into the residual block for detail enhancement. ; The residual block is specifically implemented in two ways. Convolution with padding=1 preserves the dimension. ; Output Only high-frequency detail features are enhanced, and the number of channels is the same as the input.

[0032] S3-3 Cross-layer feature fusion and channel transformation, transforming low-level small-scale features Concatenated with the low-frequency component 'a' along the channel dimension, and input into a channel transformation module consisting of two 1×1 convolutional layers, cross-layer information interaction and dimension adaptation are achieved: , , , Output The number of channels has been adjusted to , and the enhanced high-frequency components After splicing, the total number of channels is This matches the input requirements of the inverse WT transform.

[0033] S3-4 Inverse wavelet transform, feature reconstruction output: enhanced high-frequency components After channel transformation The input is concatenated by channel and fed into the inverse wavelet transform submodule. The resulting output, reconstructed using a 2×2 transposed grouped convolution, is the same size as the original large-scale features. , , Final output This feature integrates the global structure of large-scale features, high-frequency details in multiple directions, and fine-grained information of small-scale features, and can accurately adapt to the detection needs of multi-scale components of power poles.

[0034] This invention provides a preferred embodiment that executes step S4 to obtain the detection box and class probability. For example... Figure 5 As shown, it includes the following steps: The first convolutional branch of S4-1 generates and statistically optimizes the distribution of bounding box corner points, fusing features across multiple scales. Input the left branch and perform the following operations in sequence: The discrete coordinate distribution features of the output bounding box are obtained through successive 3x3 convolution, 3x3 convolution, and 1x1 convolution operations. The formula is: ; in , The maximum value is the bounding box regression value; S4-2 performs distribution statistics and quality score generation on the results of the first convolution branch: [The remaining text appears to be incomplete and requires further context.] Perform distributional statistical analysis to generate adjusted quality scores: Tensor Reshaping: Remodeling The structure; Softmax probabilistics: Calculate the probability distribution for each corner dimension. The formula is ; Top-K statistics: Extract the top 4 values ​​with the highest probability for each corner dimension. And calculate its mean. ; Feature splicing: splicing into statistical features ; MLP Transformation: Input an MLP consisting of two 1x1 convolutional layers and generate a quality score adjustment value. Furthermore, the weights and biases of the last layer of the MLP are initialized to 0 to ensure that the original score is not disturbed in the initial stage.

[0035] The second convolutional branch of S4-3 is a composite structure of depthwise separable convolution and standard convolution, which generates the initial classification score. Its structure consists of three consecutive sub-modules, and the specific process is as follows: The first submodule Perform a 3x3 depthwise separable convolution, where each channel extracts spatial features independently, and then use a 1x1 convolution to reduce the number of channels from... Adjusted to Output features ; The second submodule: (This likely refers to a specific module or component) Perform a 3x3 depthwise separable convolution to enhance spatial feature representation, followed by a 1x1 convolution to preserve the number of channels. Unchanged, output features ; The third submodule: (This likely refers to a specific module or component) Perform a 1x1 convolution to reduce the number of channels from Adjusted to the number of categories nc, the final output is the initial classification score: , in, In the table, nc=17 represents the number of categories.

[0036] S4-4: Classification Score Optimization Initial classification scores Adjusted value for quality fraction By adding each element, we obtain the optimized classification confidence score. The formula is: ,in .

[0037] S4-5 Feature stitching and detection output will display the corner point distribution features of the bounding box. Compared with the optimized classification confidence Concatenate the data along the channel dimension to obtain the input features of the detection head. , ; The final output detection boxes and category probabilities can accurately adapt to the detection needs of multi-scale components of power poles. By optimizing distributed perception, the detection confidence of small components is improved, background false detections are suppressed, and the reliability of detection results in UAV inspection scenarios is significantly enhanced.

[0038] In this embodiment, the proposed method is compared with other advanced methods on the InsPLAD dataset for UAV power pole component inspection, and commonly used evaluation metrics are selected: P, R, mAP50, Params (M), GFLOPs, Size (MB), and Inference were used. The experimental results are shown in Table 1.

[0039] Table 1 Comparison of the proposed methods on the InsPLAD dataset

[0040] As shown in Table 1, this embodiment outperforms existing methods in multiple metrics on the dataset, demonstrating the effectiveness of the method in this embodiment.

[0041] The above is a detailed description of specific embodiments of the present invention. However, the present invention is not limited to the specific embodiments described above. Those skilled in the art can make various modifications or variations to the above embodiments without affecting the essential content of the present invention, and all such modifications or variations should be within the protection scope of the present invention.

Claims

1. A method for detecting power pole components in unmanned aerial vehicle (UAV) scenarios, characterized in that, Includes the following steps: S1: Obtain a dataset of drone aerial images of power poles; S2: Preprocess the power pole and tower dataset; S3: Establish a target detection model for power pole components in a drone scenario; S3.1: In the feature extraction stage, an adaptive multi-receptive field module is adopted. The adaptive multi-receptive field module consists of a multi-expansion depth convolution sub-module, a channel attention sub-module, and a dimension transformation sub-module. The adaptive multi-receptive field module performs multi-scale feature extraction through its internal parallel multi-expansion depth convolution sub-module. This sub-module is customized for different convolution kernel sizes. By pre-setting multiple sets of convolution branches with fixed dilation rates, it covers receptive fields of different scales on the same channel dimension to enhance the ability to model the context of the target. Subsequently, feature reorganization is completed by combining the channel attention submodule and the dimension transformation submodule. The channel attention submodule is configured to adaptively generate and weight the channel weights of the input feature map. The dimension transformation submodule converts the feature tensor from batch-channel-height-width format to batch-height-width-channel format through layout transformation from NCHW to NHWC. Linear layers are used to expand the feature dimensions and introduce GELU activation. A global response normalization layer is used to normalize the global response of the expanded features to stabilize the feature distribution. Then, a linear layer is used to compress the feature dimensions. Finally, the inverse transformation from NHWC to NCHW restores the feature format to batch-channel-height-width format, and with residual connections, enhanced multi-scale features are output. S3.2: In the feature fusion stage, a hierarchical wavelet interaction unit is used to perform wavelet transform decomposition, detail enhancement, and channel interaction on adjacent scale features. The hierarchical wavelet interaction unit is configured to fuse a large-scale feature and a small-scale feature. It decomposes the large-scale feature into low-frequency components and high-frequency components in three directions (horizontal, vertical, and diagonal) through wavelet transform. The high-frequency components in the three directions are concatenated and enhanced through a residual block containing continuous 3x3 convolutions. The low-frequency components and the small-scale feature are concatenated in the channel dimension and interacted across channels through a channel transform module composed of 1x1 convolutions, compressing the feature to three times the dimension of the large-scale feature to meet the input dimension requirements of the inverse wavelet transform. The enhanced high-frequency feature is concatenated with the low-frequency small-scale feature after interaction compression and reconstructed into a fused feature of a unified scale through inverse wavelet transform, thereby achieving semantic alignment and preserving edge details in complex backgrounds. S3.3: In the prediction phase, a distribution-aware confidence optimization head is employed. This module processes the input features in parallel, generating discrete coordinate distribution features of the bounding boxes through the first convolutional branch and generating initial classification scores through the second convolutional branch. Statistical analysis is performed on the predicted bounding box distribution, and the initial classification scores are adjusted based on the Top-k probabilities and their mean statistical features, thereby optimizing the target confidence. This process uses a multilayer perceptron consisting of two 1x1 convolutional layers to perform nonlinear transformations on the statistical features, generating quality score adjustment values. The quality score adjustment values ​​are added to the initial classification scores to output the optimized classification confidence, and the final detection boxes and class probabilities are output in conjunction with the bounding box coordinate distribution features. S4: Train the model using the training set of the dataset to obtain the optimal model; S5: Use the optimal model to test the test set images of the dataset to obtain the final test results.

2. The method for detecting power pole components in a drone scenario according to claim 1, characterized in that: The InsPLAD remote sensing dataset is cropped into 640x640 image blocks without sliding window slicing or block overlap. Areas smaller than 640 pixels are zero-padded to 640x640. The label files are converted to YOLO detection format and coordinate mapping and normalization are performed based on scaling and padding relationships.

3. The method for detecting power pole components in a drone scenario according to claim 1, characterized in that: The adaptive multi-receptive field module is an improvement on the C3k2 framework, replacing the bottleneck unit with a multi-dilated depth convolution submodule. This submodule is customized for different convolution kernel sizes. When the convolution kernel size is ≥7, multi-branch dilated depth convolution is used, and its dilation rate combination is adapted according to the convolution kernel size. When the convolution kernel size is 7, a multi-branch convolution structure with dilation rates of [1,2,3] is used. For a size of 13, a dilation rate combination of [1,2,3,4,5] is used to adapt to the detection requirements of power pole components of different sizes. The channel attention submodule is configured to adaptively generate and weight the channel weights of the input feature map. It compresses the spatial dimension of the feature map through a global average pooling layer, uses convolution operations and a sigmoid activation function to generate a weight vector representing the importance of each channel, and multiplies the weight vector with the original input feature map channel by channel.

4. The method for detecting power pole components in a drone scenario according to claim 1, characterized in that: The distribution-aware confidence optimization head is configured to perform the following operations: The first convolutional branch contains two consecutive 3x3 standard convolutional layers and a final 1x1 convolutional layer, reshaping the discrete coordinate distribution features into a tensor structure of [B,M,4,H,W], where M is the maximum regression value, and processing it with the Softmax function to obtain the probability distribution of the bounding box coordinates; extracting the top K maximum probability values ​​in each coordinate dimension from the probability distribution, calculating the mean of the K probability values, and concatenating the top K maximum probability values ​​with the mean to form the distribution statistical features; the second convolutional branch adopts a hierarchical structure containing depthwise separable convolutions and standard convolutions; the value of K is 4; the hidden layer dimension of the multilayer perceptron is 64.

Images