蔡振宇,靳华伟

• 1:安徽理工大学机电工程学院

摘要(Abstract)：

针对井下管道渗漏人工巡检带来的高风险安全与成本问题，且井下光线昏暗、环境复杂，传统算法精度低与漏检误检率高等问题，提出一种基于YOLOv8算法的井下管道渗漏检测改进方法：SSO-YOLO(SPPELAN-S2-MLPv2-ODConv-YOLO).SSO-YOLO算法做出三点改进：首先对YOLOv8原有的空间金字塔池化结构，使用SPPELAN(Spatial Pyramid Pooling Efficient Layer Aggregation Network)模块改进优化；然后引入采用空间位移操作的S2-MLPv2注意力机制，改善井下环境嘈杂导致检测效果差的问题；最后，对于管道渗漏特征多变且不规则的情况，使用全维度动态卷积（ODConv）模块替换普通卷积模块，以捕捉局部多样化信息提取复杂特征．实验结果表明：模型检测实验结果的mAP50达到67.7%，相较于原模型提升了3.7个百分点．

关键词(KeyWords)： 目标检测;YOLOv8;SPPELAN;注意力机制;ODConv

Abstract:

Keywords:

基金项目(Foundation):

作者(Author): 蔡振宇,靳华伟

DOI: 10.16393/j.cnki.37-1436/z.2025.05.009

参考文献(References)：

• [1]吴谦，孙丙宇、房永峰．基于改进YOLOv5的地下排水管道缺陷检测算法[J]．阜阳师范大学学报：自然科学版，2024,41(03):21-27.
• [2]张舒，刘建文，陈泊宇．基于YOLOv7的水工输水管道目标智能识别[J]．广东水利水电，2024(09):62-67.
• [3]王治国，曹爽，管海燕，等．基于改进SSD的城市地下排水管道缺陷识别算法[J]．测绘工程，2024(05):7-13.
• [4]王合佳，林宁，林振超，等．改进YOLO的X射线管道焊缝检测算法[J]．华侨大学学报：自然科学版，2024,45(06):766-775.
• [5]YAN S,CHANG J,LI L,et al.Improved Object Detection Network for Pipeline Leakage Localization in the Pneumatic System Based on Thermal Images[J].Measurement,2025,242:116225-116234.
• [6]MICHAEL G,SHAHRA E Q,BASURR S,et al.Real-Time Pipeline Fault Detection in Water Distribution Networks Using You Only Look Once v8[J].Sensors,2024,24(21):6982-6992.
• [7]YANG D,MA C,YU G,et al.Automatic Defect Detection of Pipelines Based on Improved OFG-YOLO Algorithm[J].Measurement,2025,242:115847-115856.
• [8]G R D,B P.YoloTransformer-TransDetect:A Bybrid Model for Steel Tube Defect Detection Using YOLO and Transformer Architectures[J].International Journal on Interactive Design and Manufacturing(IJIDeM),2024:1-17.
• [9]ZHANG S,HU K,SHA W,et al.Efficient One-Stage Location Method for Grape Picking Points in Natural Scene by Combining Detection Network and Point Regression[J].Computers and Electronics in Agriculture,2025,230:109725-109734.
• [10]WANG D,TAN J,WANG H,et al.SDS-YOLO:An Improved Vibratory Position Detection Algorithm Based on YOLOv11[J].Measurement,2025,244:116518-116527.
• [11]ISLAM U Q,KHOZAEI F,BARHOUMI A S M E,et al.Advancing Autonomous SLAM Systems:Integrating YOLO Object Detection and Enhanced Loop Closure Techniques for Robust Environment Mapping[J].Robotics and Autonomous Systems,2025,185:104871-104880.
• [12]SHANG D,LV Z,GAO Z,et al.Detection of Coal Gangue by YOLO Deep Learning Method Based on Channel Pruning[J].International Journal of Coal Preparation and Utilization,2025,45(01):231-240.
• [13]GAO F,YANG Z,LIU Q,et al.Enhanced Oracle Bone Corrosion Detection Using Attention-Guided YOLO With Ghost Convolution[J].Discover Artificial Intelligence,2024,4(01):109-117.
• [14]SHEN Q,ZHANG X,SHEN M,et al.Multi-Scale Adaptive YOLO for Instance Segmentation of Grape Pedicels[J].Computers and Electronics in Agriculture,2025,229:109712-109721.
• [15]JI P,WU C,ZHANG X,et al.A YOLO Network Based on Depthwise Convolution Attention,Feature Fusion,and KL Divergence(DFK-YOLO):A Deep Learning Method for Infrared Small Target Detection Based on YOLOv7[J].Electronics,2024,13(23):4820-4830.
• [16]SHUANG L,JIAN Ye H,Chenxiang L,et al.Application of automatic structured pruning method based on YOLO in power operation scenarios[J].Intelligent Decision Technologies,2024,18(04):3323-3332.
• [17]XU G,LI A,WANG X,et al.Research on X-Ray Nondestructive Defect Detection Method of Tire Based on Dynamic Snake Convolution YOLO Model[J].Scientific Reports,2024,14(01):29587-29596.
• [18]XU J,ZHAO L,REN Y,et al.LightYOLO:Lightweight Model Based on YOLOv8nfor Defect Detection of Ultrasonically Welded Wire Terminations[J].Engineering Science and Technology,an International Journal,2024,60:101896-101905.