基于自适应特征提取的通信光缆缺陷检测方法

doi:10.11996/JG.j.2095-302X.2025020241

摘要/Abstract

摘要：

随着通信线路覆盖面积的增长，传统全介质自承式光缆(ADSS)电腐蚀缺陷巡检方式存在效率低、成本高等缺点，据此提出一种基于特征自适应提取的ADSS通信光缆电腐蚀缺陷检测方法，通过对YOLOv8n模型进行针对性改进，实现对ADSS光缆电腐蚀缺陷图像的检测。首先在主干网络中引入ADown下采样模块，在下采样过程中保留更多光缆特征的细节信息。随后引入上下文特征增强模块，使算法更有针对性地学习光缆缺陷特征。最后提出一种基于特征自适应提取的C2f_DSC模块，在颈部网络中加入，利用动态蛇形卷积特性加强对光缆区域特征的提取。在ADSS光缆电腐蚀缺陷数据集上进行实验，改进算法相比基线模型YOLOv8n在mAP50精度上提高了2.5%，在mAP50∶95精度上提高了2.2%，为ADSS光缆巡检提供了一种新的有效方法。

关键词: ADSS光缆, 电腐蚀, 缺陷检测, 特征自适应提取, YOLO

Abstract:

With the expansion of communication line coverage, the traditional inspection method for electrical corrosion defects in all dielectric self-supporting (ADSS) optical cables have faced issues of low efficiency and high costs. To address these issues, a detection method for electrical corrosion defects in ADSS communication cables based on adaptive feature extraction was proposed. This method achieved detection of electrical corrosion defects in ADSS cables by making targeted improvements to the YOLOv8n model. Firstly, an ADown downsampling module was introduced into the backbone network to preserve more detailed information about the cable during the downsampling process. Subsequently, a context feature enhancement module was introduced, enabling the algorithm to learn the defect features of optical cables more specifically. Finally, a C2f_DSC module based on feature adaptive extraction was proposed, utilizing the dynamic serpentine convolution feature in the neck network to enhance the extraction of cable area features. Experiments conducted on an ADSS cable electrical corrosion dataset demonstrated that compared to the baseline model YOLOv8n, the proposed algorithm achieved a 2.5% improvement in mAP50 accuracy and a 2.2% increase in mAP50∶95 accuracy, providing a new and effective method for ADSS cable inspection.

Key words: ADSS optic cable, electrical corrosion, defect detection, feature adaptive extraction, YOLO

中图分类号:

TP391
TN913

王志东, 陈晨阳, 刘晓明. 基于自适应特征提取的通信光缆缺陷检测方法[J]. 图学学报, 2025, 46(2): 241-248.

WANG Zhidong, CHEN Chenyang, LIU Xiaoming. Defect detection method of communication optical cable based on adaptive feature extraction[J]. Journal of Graphics, 2025, 46(2): 241-248.

图/表 10

图1 改进模型网络结构图

Fig. 1 Improved model network structure diagram

图2 ADown模块结构图

Fig. 2 ADown structure diagram

图3 CAA结构图

Fig. 3 CAA structure diagram

图4 动态蛇形卷积网格变化图

Fig. 4 Mesh variation diagram for dynamic snake convolution

图5 C2f_DSC结构图

Fig. 5 C2f_DSC structure diagram

图6 ADSS光缆电腐蚀缺陷图((a)击穿；(b)断缆；(c)坑蚀；(d)局部坑蚀；(e)电痕；(f)背景)

Fig. 6 Diagram of electrical corrosion in ADSS cable ((a) Breakdown; (b) Break; (c) Pitting; (d) Smallpitting; (e) Tracking; (f) Background)

表1 消融实验结果

Table 1 Ablation of experiment results

模型	P/%	R/%	mAP50/%	mAP50∶95/%	Params/M	GFLOPs
YOLOv8n	84.6	80.2	85.7	49.7	3.01	8.2
YOLOv8s+A	87.3	82.3	86.9	50.6	2.61	7.7
YOLOv8s+A+B	88.4	81.3	87.4	50.8	2.75	7.8
YOLOv8s+A+B+C	88.8	82.6	88.2	51.9	3.17	8.4

图7 验证损失变化

Fig. 7 Verification of loss variation

表2 对比实验结果

Table 2 Comparison of experiment results

模型	mAP50/%	mAP50∶95/%	Params/M	GFLOPs	Time/ms	mAP50/%
模型	mAP50/%	mAP50∶95/%	Params/M	GFLOPs	Time/ms	断缆	击穿	坑蚀	电痕	局部坑蚀
SSD	77.1	33.3	24.28	30.63	31.40	89.8	75.2	69.3	73.4	77.9
Faster R-CNN	81.6	37.6	60.36	247.81	38.50	92.1	69.8	75.4	84.3	86.3
YOLOv3-tiny	84.3	45.9	12.13	19.00	1.86	92.6	88.6	77.4	79.6	83.3
YOLOv5n	85.1	48.3	2.51	7.20	4.66	92.2	92.5	74.8	79.2	86.7
YOLOv6n	85.1	49.5	4.24	11.90	3.84	89.4	88.5	81.7	79.6	86.0
YOLOv7-tiny	85.2	45.7	6.03	13.20	4.46	94.3	91.2	77.8	79.7	83.2
RT-DETR-R18	87.7	51.6	20.08	58.30	22.39	93.1	91.8	81.2	84.6	87.6
YOLOv10n	83.0	47.0	2.30	6.70	3.62	94.3	87.6	74.4	77.3	81.2
YOLOv8n	85.7	49.7	3.01	8.20	4.31	92.9	91.4	76.6	81.0	86.5
Ours	88.2	51.9	3.17	8.40	5.09	96.3	92.4	83.3	83.1	86.2

图8 光缆缺陷检测效果对比图((a)原图；(b) RT-DETR；(c) YOLOv8n；(d) YOLOv10n；(e)本文算法)

Fig. 8 Comparison of optical cable defect detection effects ((a) Original; (b) RT-DETR; (c) YOLOv8n; (d) YOLOv10n; (e) Ours)

参考文献 17

[1]	陈少磊, 罗洋, 邓平, 等. ADSS光缆电腐蚀机理研究及防范措施应用[J]. 电力信息与通信技术, 2019, 17(9): 67-73.
	CHEN S L, LUO Y, DENG P, et al. Research on electro-corrosion mechanism of ADSs optical cable and application of preventive measures[J]. Electric Power Information and Communication Technology, 2019, 17(9): 67-73 (in Chinese).
[2]	王富. ADSS光缆电腐蚀故障的在线监测措施[J]. 电子世界, 2018(14): 171, 173.
	WANG F. Online monitoring measures for galvanic corrosion faults in ADSS fiber optic cables[J]. Electronics World, 2018(14): 171, 173 (in Chinese).
[3]	刘传洋, 吴一全. 基于深度学习的输电线路视觉检测方法研究进展[J]. 中国电机工程学报, 2023, 43(19): 7423-7445.
	LIU C Y, WU Y Q. Research progress of vision detection methods based on deep learning for transmission lines[J]. Proceedings of the CSEE, 2023, 43(19): 7423-7445 (in Chinese).
[4]	GIRSHICK R, DONAHUE J, DARRELL T, et al. Rich feature hierarchies for accurate object detection and semantic segmentation[C]// 2014 IEEE Conference on Computer Vision and Pattern Recognition. New York: IEEE Press, 2014: 580-587.
[5]	LIU W, ANGUELOV D, ERHAN D, et al. SSD: single shot MultiBox detector[C]// The 14th European Conference on Computer Vision. Cham: Springer, 2016: 21-37.
[6]	REDMON J, DIVVALA S, GIRSHICK R, et al. You only look once: unified, real-time object detection[C]// 2016 IEEE Conference on Computer Vision and Pattern Recognition. New York: IEEE Press, 2016: 779-788.
[7]	裴少通, 张行远, 胡晨龙, 等. 基于ER-YOLO算法的跨环境输电线路缺陷识别方法[J]. 电工技术学报, 2024, 39(9): 2825-2840.
	PEI S T, ZHANG H X, HU C L, et al. The defect detection method for cross-environment power transmission line based on ER-YOLO algorithm[J]. Transactions of China Electrotechnical Society, 2024, 39(9): 2825-2840 (in Chinese).
[8]	王道累, 康博, 朱瑞. 基于深度学习的电力设备铭牌文本检测方法[J]. 图学学报, 2023, 44(4): 691-698. DOI
	WANG D L, KANG B, ZHU R. Text detection method for electrical equipment nameplates based on deep learning[J]. Journal of Graphics, 2023, 44(4): 691-698 (in Chinese).
[9]	WANG C Y, LIAO H Y M, WU Y H, et al. CSPNet: a new backbone that can enhance learning capability of CNN[C]// 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops. New York: IEEE Press, 2020: 1571-1580.
[10]	LIU S, QI L, QIN H F, et al. Path aggregation network for instance segmentation[C]// 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. New York: IEEE Press, 2018: 8759-8768.
[11]	LIN T Y, DOLLÁR P, GIRSHICK R, et al. Feature pyramid networks for object detection[C]// 2017 IEEE Conference on Computer Vision and Pattern Recognition. New York: IEEE Press, 2017: 936-944.
[12]	WANG C Y, YEH I H, LIAO H Y M. YOLOv9:learning what you want to learn using programmable gradient information[EB/OL]. [2024-06-13]. https://arxiv.org/abs/2402.13616.
[13]	CAI X H, LAI Q X, WANG Y W, et al. Poly kernel inception network for remote sensing detection[C]// 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition. New York: IEEE Press, 2024: 27706-27716.
[14]	QI Y L, HE Y T, QI X M, et al. Dynamic snake convolution based on topological geometric constraints for tubular structure segmentation[C]// 2023 IEEE/CVF International Conference on Computer Vision. New York: IEEE Press, 2023: 6047-6056.
[15]	ZHAO Y A, LV W Y, XU S L, et al. DETRs beat YOLOs on real-time object detection[C]// 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition. New York: IEEE Press, 2024: 16965-16974.
[16]	曹义亲, 伍铭林, 徐露. 基于改进YOLOv5算法的钢材表面缺陷检测[J]. 图学学报, 2023, 44(2): 335-345. DOI
	CAO Y Q, WU M L, XU L. Steel surface defect detection based on improved YOLOv5 algorithm[J]. Journal of Graphics, 2023, 44(2): 335-345 (in Chinese).
[17]	WANG A, CHEN H, LIU L H, et al. YOLOv10:Real-time end-to-end object detection[EB/OL]. [2024-06-13]. https://arxiv.org/abs/2405.14458.