《计算机应用》唯一官方网站 ›› 2026, Vol. 46 ›› Issue (9): 2987-2995.DOI: 10.11772/j.issn.1001-9081.2025080989

• 多媒体计算与计算机仿真 • 上一篇    

轻量级融合反射感知特征的道路积水检测网络REL-YOLO

闫奥运1, 朱灵龙2,3()   

  1. 1.南京信息工程大学 计算机学院,南京 210044
    2.无锡学院 物联网工程学院,江苏 无锡 214105
    3.公安部交通管理科学研究所,江苏 无锡 214151
  • 收稿日期:2025-08-27 修回日期:2025-11-18 接受日期:2025-11-27 发布日期:2025-12-01 出版日期:2026-09-10
  • 通讯作者: 朱灵龙
  • 作者简介:闫奥运(2001—),男,安徽宿州人,硕士研究生,主要研究方向:计算机视觉
    朱灵龙(1993—),男,江苏启东人,副教授,博士,主要研究方向:交通大数据建模、人工智能。
  • 基金资助:
    国家自然科学基金资助项目(42305158);国家自然科学基金资助项目(42105143);无锡市“太湖之光”基础研究项目(K20231021);江苏省高等学校基础科学(自然科学)研究面上项目(23KJB170025)

REL-YOLO: lightweight road waterlogging detection network integrating reflection perception features

Aoyun YAN1, Linglong ZHU2,3()   

  1. 1.School of Computer Science,Nanjing University of Information Science and Technology,Nanjing Jiangsu 210044,China
    2.School of Internet of Things Engineering,Wuxi University,Wuxi Jiangsu 214105,China
    3.Traffic Management Research Institute of the Ministry of Public Security,Wuxi Jiangsu 214151,China
  • Received:2025-08-27 Revised:2025-11-18 Accepted:2025-11-27 Online:2025-12-01 Published:2026-09-10
  • Contact: Linglong ZHU
  • About author:YAN Aoyun, born in 2001, M. S. candidate. His research interests include computer vision.
    ZHU Linglong, born in 1993, Ph. D., associate professor. His research interests include traffic big data modeling, artificial intelligence.
  • Supported by:
    National Natural Science Foundation of China(42305158);Wuxi “Taihu Light” Basic Research Project(K20231021);Jiangsu Provincial Higher Education Basic Science (Natural Science) Research General Project(23KJB170025)

摘要:

道路积水作为一种广泛存在的交通安全隐患,易导致车辆失控,引发交通事故。针对准确识别道路积水受限于复杂路面纹理、多变光照条件及积水本身形态多样等挑战,基于YOLOv11n,提出轻量级融合反射感知特征的道路积水检测网络REL-YOLO(Reflection-aware Enhanced Lightweight YOLO)。首先,设计反射感知特征增强(RAFE)模块,以有效融合浅层纹理与深层语义特征,并通过自适应噪声加权机制抑制镜面反射干扰;其次,设计C3k2_Enhanced模块,结合分组卷积与轻量化通道注意力机制,在降低参数量的同时增强对细小水流特征的捕获能力;最后,引入轻量级共享细节增强检测头(Detect_LSDECD),通过共享卷积与可学习尺度参数优化多尺度特征融合。实验结果表明,与原始的YOLOv11n模型相比,REL-YOLO在Road waterlogging数据集上的召回率和mAP@50分别提升了3.2和2.0个百分点,同时模型参数量减少了13.1%,浮点运算量由6.3 GFLOPs降低至4.3 GFLOPs。REL-YOLO在保持高实时性的同时,实现了精度与效率的平衡。

关键词: 道路积水, 目标检测, 反射感知, 特征增强, 轻量化模型

Abstract:

Road waterlogging is regarded as a widespread traffic safety hazard that can easily lead to vehicle loss of control and then trigger traffic accidents. Accurate identification of road waterlogging is limited by challenges such as complex road surface textures, variable lighting conditions, and the diverse morphology of puddle itself. To address these issues, REL-YOLO (Reflection-aware Enhanced Lightweight YOLO), a lightweight road waterlogging detection network integrating reflection perception features, was proposed on the basis of YOLOv11n. First, the Reflection-Aware Feature Enhancement (RAFE) module was designed to fuse shallow texture and deep semantic features effectively, and specular reflection interference was suppressed through an adaptive noise weighting mechanism. Second, the C3k2_Enhanced module was designed by combining grouped convolution with a lightweight channel attention mechanism, so that the number of parameters was reduced while the capture capability of fine water flow features was enhanced. Finally, a Lightweight Shared Detail Enhanced Convolutional Detection head (Detect_LSDECD) was introduced, and multi-scale feature fusion was optimized through shared convolution and learnable scale parameters. Experimental results show that compared with the original YOLOv11n model on the Road waterlogging dataset, REL-YOLO has recall improved by 3.2 percentage points, mAP@50 improved by 2.0 percentage points, the number of model parameters reduced by 13.1%, and floating-point operations reduced from 6.3 GFLOPs to 4.3 GFLOPs. While maintaining high real-time performance, REL-YOLO achieves a balance between accuracy and efficiency.

Key words: road waterlogging, object detection, reflection perception, feature enhancement, lightweight model

中图分类号: