基于改进YOLOv5s的输电塔螺栓松动检测

doi:10.12141/j.issn.1000-565X.250172

摘要/Abstract

摘要：

输电塔作为电力输送网络的关键基础设施，其结构安全性直接关系到电网的稳定运行。在长期服役过程中，输电塔螺栓受风荷载、温差效应及材料老化等多因素耦合作用，易逐渐发生松动。该文提出了一种基于改进YOLOv5s的输电塔螺栓松动智能检测模型（CCSGS-YOLO）：采用坐标卷积替代主干网络中的标准卷积层，增强模型对目标位置信息的获取能力；引入卷积注意力模块（CBAM），通过通道与空间双重注意力机制，强化模型在复杂背景下的特征鉴别能力；构建Slim-Neck特征融合结构，借助跨阶段部分连接与深度可分离卷积的优化组合，在维持检测精度的同时降低计算复杂度；采用GIoU损失函数与Soft-NMS的联合优化策略，通过考虑预测框与真实框的重叠几何特性，提升目标检测的定位精度。实验结果表明：CCSGS-YOLO的精确率达91.7%，召回率为89.4%，平均精度均值达到95.3%，F₁分数提升至90.0%，较基准模型YOLOv5s分别提高了1.6、3.0、1.4和1.0个百分点；在计算效率方面，CCSGS-YOLO模型检测速度达74.8 f/s，推理时延降低至13.4 ms，较YOLOv5s模型提升11.6%。此外，该文通过现场实验验证了CCSGS-YOLO在不同场景下的检测鲁棒性，为输电塔螺栓松动的智能巡检提供了一种新思路。

关键词: 输电塔, 螺栓松动, 深度学习, 目标检测

Abstract:

As the critical infrastructure in power transmission networks, the structural stability of transmission towers directly impacts the safe and reliable operation of the power grid. During long-term service, bolts in the tower structure are prone to gradual loosening under the coupled effects of multiple factors such as wind loads, temperature variations, and material aging. This paper proposed an intelligent detection model for bolt looseness in transmission towers based on an improved YOLOv5s (named CCSGS-YOLO). The model incorporates several key enhancements: coordinate convolution replaces standard convolutional layers in the backbone network to strengthen the model’s ability to capture positional information of targets; a convolutional block attention module (CBAM) is introduced to strengthen the model’s feature discrimination capability in complex backgrounds through dual channel and spatial attention mechanisms; a slim-neck feature fusion architecture is constructed, leveraging an optimized combination of cross-stage partial connections and depthwise separable convolutions to reduce computational complexity while maintaining detection accuracy; a joint optimization strategy employing the Generalized Intersection over Union (GIoU) loss function and Soft Non-Maximum Suppression (Soft-NMS) improves localization accuracy by considering the geometric overlap characteristics between predicted and ground-truth bounding boxes. Experimental results demonstrate that CCSGS-YOLO achieves a precision of 91.7%, a recall of 89.4%, a mean average precision (mAP) of 95.3%, and an F₁ score of 90.0%. These metrics represent improvements of 1.6, 3.0, 1.4, and 1.0 percentage points, respectively, over the baseline YOLOv5s model. In terms of computational efficiency, the model achieves a detection speed of 74.8 frames per second (FPS), reducing the inference latency to 13.4 ms, which represents an 11.6% improvement compared to the YOLOv5s model. Furthermore, this paper validates the detection robustness of CCSGS-YOLO across various scenarios through field experiments, providing a novel approach for intelligent inspection of loose bolts on transmission towers.

Key words: transmission tower, bolt looseness, deep learning, target detection

中图分类号:

TP391.4

王德弘, 张子轩. 基于改进YOLOv5s的输电塔螺栓松动检测[J]. 华南理工大学学报(自然科学版), 2026, 54(2): 25-37.

WANG Dehong, ZHANG Zixuan. Transmission Tower Bolt Looseness Detection Based on Improved YOLOv5s[J]. Journal of South China University of Technology(Natural Science Edition), 2026, 54(2): 25-37.

图/表 18

图1

图2

图3

图4

图5

图6

图 7

图8

图9

图10

图11

图12

表 2

图13

表 3

图14

图15

图16

参考文献 27

[1]	WANG D， HOU S， RUAN Y，et al ．Research on the slip model of main member bar connection of transmission tower using bolts［J］．International Journal of Steel Structures，2025，25（1）：68-79.
[2]	杨风利．考虑螺栓连接滑移影响的输电铁塔塔身结构分析［J］．工程力学，2018，35（）：193-199.
	YANG Feng-li ．Structural analysis on a typical transmission tower body section with bolt slippage effects［J］．Engineering Mechanics，2018，35（）：193-199.
[3]	李嘉祥，张超，程金鹏，等．输电塔螺栓搭接节点滞回性能试验研究［J］．振动与冲击，2023，42（22）：10-18，120.
	LI Jiaxiang， ZHANG Chao， CHENG Jinpeng，et al ．Experimental study on the hysteresis performance of bolted lap joints of a transmission tower［J］．Journal of Vibration and Shock，2023，42（22）：10-18，120.
[4]	HUO L， CHEN D， LIANG Y，et al ．Impedance based bolt pre-load monitoring using piezoceramic smart washer［J］．Smart Materials and Structures，2017，26（5）：057004/1-7.
[5]	WU G， XU C， DU F，et al ．A modified time reversal method for guided wave detection of bolt loosening in simulated thermal protection system panels［J］．Complexity，2018，2018：8210817/1-12.
[6]	ZHOU L， CHEN S X， NI Y Q，et al ．EMI-GCN：a hybrid model for real-time monitoring of multiple bolt looseness using electromechanical impedance and graph convolutional networks［J］．Smart Materials and Structures，2021，30（3）：035032/1-20.
[7]	ZHANG Y， ZHAO X， SUN X，et al ．Bolt loosening detection based on audio classification［J］．Advances in Structural Engineering，2019，22（13）：2882-2891.
[8]	WANG X， YUE Q， LIU X ．Bolted lap joint loosening monitoring and damage identification based on acoustic emission and machine learning［J］．Mechanical Systems and Signal Processing，2024，220：111690/1-15.
[9]	JIANG Z， ZHU Z， ZHUO D ．Identification of bolt loosening damage of steel truss structure based on MFCC-WPES and optimized random forest［J］．Applied Sciences，2024，14（15）：6626/1-18.
[10]	WANG F， SONG G ．A novel percussion-based method for multi-bolt looseness detection using one-dimensional memory augmented convolutional long short-term memory networks［J］．Mechanical Systems and Signal Processing，2021，161：107955/1-12.
[11]	WANG S， ZHOU Y， KONG Q ．A force-adaptive percussion method for bolt looseness assessment［J］．Journal of Civil Structural Health Monitoring，2024，14（4）：831-841.
[12]	张姝，王昊天，董骁翀，等．基于深度学习的输电线路螺栓检测技术［J］．电网技术，2021，45（7）：2821-2828.
	ZHANG Shu， WANG Haotian， DONG Xiaochong，et al ．Bolt detection technology of transmission lines based on deep learning［J］．Power System Technology，2021，45（7）：2821-2828.
[13]	SUN Y， LI M， DONG R，et al ．Vision-based detection of bolt loosening using YOLOv5［J］．Sensors，2022，22（14）：5184/1-16.
[14]	PENG L， WANG K， ZHOU H，et al ．YOLOv7-CWFD for real time detection of bolt defects on transmission lines［J］．Scientific Reports，2025，15（1）：1635/1-17.
[15]	HUA G， ZHANG H， HUANG C，et al ．An enhanced YOLOv8‐based bolt detection algorithm for transmission line［J］．IET Generation，Transmission & Distribution，2024，18（24）：4065-4077.
[16]	CHEN Z， WANG L， LI B，et al ．An improved faster R-CNN transmission line bolt defect detection method［C］∥ Proceedings of 2022 World Automation Congress.San Antonio：IEEE，2022：82-85.
[17]	WU T， SHANG K， DAI W，et al ．High-resolution cross-scale transformer：a deep learning model for bolt loosening detection based on monocular vision measurement［J］．Engineering Applications of Artificial Intelligence，2024，133：108574/1-15.
[18]	ZHANG Y， SUN X， LOH K J，et al ．Autonomous bolt loosening detection using deep learning［J］．Structural Health Monitoring，2020，19（1）：105-122.
[19]	ZHANG Y， YUEN K V ．Bolt damage identification based on orientation-aware center point estimation network［J］．Structural Health Monitoring，2022，21（2）：438-450.
[20]	邢岩，郭思豪，张振，等．基于CGT-YOLO的小目标交通标志识别算法［J］．华南理工大学学报（自然科学版），2025-11-25，doi：10.12141/j.issn.1000-565X.250092 .
	XING Yan， GUO Sihao， ZHANG Zhen，et al ．Small traffic sign recognition algorithm based on CGT-YOLO［J］．Journal of South China University of Technology（Natural Science Edition），2025-11-25，doi：10.12141/j.issn.1000-565X.250092 .
[21]	LIU R， LEHMAN J， MOLINO P，et al ．An intriguing failing of convolutional neural networks and the CoordConv solution［C］∥ Proceedings of the 32nd International Conference on Neural Information Processing Systems．Red Hook：ACM，2018：9628-9639.
[22]	WOO S， PARK J， LEE J Y，et al ．CBAM： convolutional block attention module［C］∥ Proceedings of the 15th European Conference on Computer Vision．Munich：Springer，2018：3-19.
[23]	LI H， LI J， WEI H，et al ．Slim-neck by GSConv：a lightweight-design for real-time detector architectures［J］．Journal of Real-Time Image Processing，2024，21（3）：62/1-13.
[24]	YU J， JIANG Y， WANG Z，et al ．UnitBox：an advanced object detection network［C］∥ Proceedings of the 24th ACM International Conference on Multimedia．New York：ACM，2016：516-520.
[25]	REZATOFIGHI H， TSOI N， GWAK J Y，et al ．Generalized intersection over union：a metric and a loss for bounding box regression［C］∥ Proceedings of 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition．Long Beach：IEEE，2019：658-666.
[26]	BODLA N， SINGH B， CHELLAPPA R，et al ．Soft-NMS：improving object detection with one line of code［C］∥ Proceedings of 2017 IEEE International Conference on Computer Vision．Venice：IEEE，2017：5562-5570.
[27]	EVERINGHAM M， GOOL L， WILLIAMS I K C，et al ．The Pascal visual object classes （VOC） challenge［J］．International Journal of Computer Vision，2010，88（2）：303-338.

CoordConv	CBAM	Slim-Neck	GIoU+Soft-NMS	P/%	R/%	mAP@0.5/%	参数量/10⁶	浮点运算次数/10⁹
×	×	×	×	90.1	86.4	93.9	7.02	16.0
√	×	×	×	90.5	87.5	94.3	7.02	16.0
×	√	×	×	92.5	89.0	95.0	7.03	16.1
×	×	√	×	92.3	87.7	94.6	4.23	13.1
×	×	×	√	91.3	88.7	94.7	7.02	16.0
√	√	√	√	91.7	89.4	95.3	4.24	13.2

模型	P/%	R/%	mAP@0.5/%	F₁/%	检测速度/（f‧s^-1）
YOLOv3-Tiny	85.7	79.3	87.3	82.0	65.6
YOLOv5s	90.1	86.4	93.9	89.0	67.0
YOLOv7-Tiny	89.4	87.8	94.0	89.0	69.4
YOLOv8n	90.7	88.0	94.3	88.0	72.2
CCSGS-YOLO	91.7	89.4	95.3	90.0	74.8