[1]汪杰,梁平,廖光忠.基于YOLOv8改进的道路缺陷检测算法[J].计算机技术与发展,2025,(10):35-42.[doi:10.20165/j.cnki.ISSN1673-629X.2025.0141]
　WANG Jie,LIANG Ping,LIAO Guang-zhong.Improved Road Defect Detection Algorithm Based on YOLOv8[J].,2025,(10):35-42.[doi:10.20165/j.cnki.ISSN1673-629X.2025.0141]
点击复制

基于YOLOv8改进的道路缺陷检测算法()

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:
期数:
2025年10期
页码:
35-42
栏目:
媒体计算
出版日期:
2025-10-10

文章信息/Info

Title:
Improved Road Defect Detection Algorithm Based on YOLOv8
文章编号:
1673-629X(2025)10-0035-08
作者:
汪杰1; 2; 梁平1; 2; 廖光忠1; 2
1. 武汉科技大学 计算机科学与技术学院,湖北武汉 430065;
2. 武汉科技大学 智能信息处理与实时工业系统湖北省重点实验室,湖北武汉 430065
Author(s):
WANG Jie1; 2; LIANG Ping1; 2; LIAO Guang-zhong1; 2
1. School of Computer Science and Technology,Wuhan University of Science and Technology,Wuhan 430065,China;
2. Hubei Province Key Laboratory of Intelligent Information Processing and Real-time Industrial System,Wuhan University of Science and Technology,Wuhan 430065,China
关键词:
目标检测; YOLOv8; 道路缺陷检测; 注意力机制; 动态检测头
Keywords:
object detection; YOLOv8; road defect detection; attention mechanism; dynamic detection head
分类号:
TP391.4
DOI:
10.20165/j.cnki.ISSN1673-629X.2025.0141
摘要:
针对道路缺陷检测中轻量化与检测精度难以兼顾的问题,该文在YOLOv8模型基础上提出了一种平衡模型IMD-YOLOv8。首先,通过融合倒置残差结构iRMB与高效多尺度注意力EMA,设计C2f-iEMA模块替换原C2f模块,在保持参数量相近的前提下,利用跨尺度特征重组增强病害特征提取能力,显著提升小目标检测精度;其次,引入多尺度空洞注意力MSDA优化局部特征交互,通过多尺度语义聚合进一步提升检测鲁棒性,并采用动态检测头DynamicHead替换固定结构,通过任务感知注意力动态分配计算资源,显著减少冗余参数;最后,设计形状敏感的ShapeIoU损失函数,通过约束边界框形状相似性优化回归精度,避免引入额外计算开销。实验表明,IMD-YOLOv8在RDD2022数据集上实现mAP@0.5为89.4%,较YOLOv8提升2.9百分点,参数量与计算量分别减少20百分点和3百分点,FPS达98.5f/s,在轻量化与精度之间实现有效平衡。
Abstract:
Aiming at the contradiction between lightweight and detection accuracy in road defect detection,we propose a balanced model IMD-YOLOv8 based on the YOLOv8 model. Firstly,by fusing the inversion residual structure iRMB and efficient multi-scale attention EMA,the C2F-IEMA module was designed to replace the original C2f module. Under the premise of maintaining a similar number of parameters,the cross-scale feature recombination was used to enhance the ability of disease feature extraction and significantly improve the accuracy of small target detection. Secondly,the multi-scale hole attention MSDA was introduced to optimize the local feature interaction,and the detection robustness was further improved through multi-scale semantic aggregation. The fixed structure was replaced by DynamicHead,and the task-aware attention was used to dynamically allocate computing resources,which significantly reduced redundant parameters. Finally,the shape-sensitive ShapeIoU loss function is designed to optimize the regression accuracy by restricting the shape similarity of the bounding box to avoid the introduction of additional computational overhead. Experiments show that IMD-YOLOv8 achieves mAP@0.5 of 89.4% on the RDD2022 dataset,which is 2.9 percentage point higher than that of YOLOv8,reduces the amount of parameters and calculation by 20 percentage point and 3 percentage point respectively,and achieves a FPS of 98.5 f/s,which achieves an effective balance between lightweight and accuracy.

相似文献/References:

[1]刘晓明 李毓蕙 高燕 郑华强.基于目标区域清晰显示的H.264编码策略[J].计算机技术与发展,2010,(06):29.
　LIU Xiao-ming,LI Yu-hui,GAO Yan,et al.A Coding Strategy of H.264 Based on High-definition Display of Target Region[J].,2010,(10):29.
[2]刘翔 吴谨 祝愿博 康晓晶.基于视频序列的目标检测与跟踪技术研究[J].计算机技术与发展,2009,(11):179.
　LIU Xiang,WU Jin,ZHU Yuan-bo,et al.A Study of Object Detecting and Tracking Based on Video Sequences[J].,2009,(10):179.
[3]曙光 张超 蔡则苏.基于改进的混合高斯模型的目标检测方法[J].计算机技术与发展,2012,(07):60.
　SHU Guang,ZHANG Chao,CAI Ze-su.Target Detection Method Based on Improved Gaussian Mixture Model[J].,2012,(10):60.
[4]刘洁,李目,周少武.一种混沌混合粒子群优化RBF神经网络算法[J].计算机技术与发展,2013,(08):181.
　LIU Jie[],LI Mu[],ZHOU Shao-wu[].An Algorithm of Chaotic Hybrid Particle Swarm Optimization Based on RBF Neural Network[J].,2013,(10):181.
[5]蒋翠清,孙富亮,吴艿芯. 基于相对欧氏距离的背景差值法视频目标检测[J].计算机技术与发展,2015,25(01):37.
　JIANG Cui-qing,SUN Fu-liang,WU Nai-xin. Video Object Detection of Background Subtraction Method Based on Relative Euclidean Distance[J].,2015,25(10):37.
[6]卢官明,衣美佳. 步态识别关键技术研究[J].计算机技术与发展,2015,25(07):100.
　LU Guan-ming,YI Mei-jia. Research on Critical Techniques in Gait Recognition[J].,2015,25(10):100.
[7]高翔,朱婷婷,刘洋. 多摄像头系统的目标检测与跟踪方法研究[J].计算机技术与发展,2015,25(07):221.
　GAO Xiang,ZHU Ting-ting,LIU Yang. Research of Target Detection and Tracking Method for Multi-camera System[J].,2015,25(10):221.
[8]章文洁[][],黄旻[],张桂峰[]. 滤光片多光谱成像中运动目标场景误配准修正[J].计算机技术与发展,2016,26(01):18.
　ZHANG Wen-jie[][],HUANG Min[],ZHANG Gui-feng[]. Misregistration Correction for Moving Object Scene in Filter-type Multispectral Imaging[J].,2016,26(10):18.
[9]施泽浩,赵启军.基于全卷积网络的目标检测算法[J].计算机技术与发展,2018,28(05):55.[doi:10.3969/j.issn.1673－629X.2018.05.013]
　SHI Ze-hao,ZHAO Qi-jun.Object Detection Algorithm Based on Fully Convolutional Neural Network[J].,2018,28(10):55.[doi:10.3969/j.issn.1673－629X.2018.05.013]
[10]张夏清,茅耀斌. 一种改进的ViBe背景提取算法[J].计算机技术与发展,2016,26(07):36.
　ZHANG Xia-qing,MAO Yao-bin. An Improved ViBe Background Generation Method[J].,2016,26(10):36.
[11]高雪豪,吴建平,韦杰,等.基于增强多尺度融合YOLOv8的道路病害检测算法[J].计算机技术与发展,2024,34(11):140.[doi:10.20165/j.cnki.ISSN1673-629X.2024.0217]
　GAO Xue-hao,WU Jian-ping,WEI Jie,et al.Road Disease Detection Algorithm Based on Enhanced Multi-scale Fusion YOLOv8[J].,2024,34(10):140.[doi:10.20165/j.cnki.ISSN1673-629X.2024.0217]
[12]赵龙阳,李天豪,张会兵,等.MDSD-YOLO:一种复杂道路场景目标检测方法[J].计算机技术与发展,2025,(09):30.[doi:10.20165/j.cnki.ISSN1673-629X.2025.0113]
　ZHAO Long-yang,LI Tian-hao,ZHANG Hui-bing,et al.MDSD-YOLO:An Object Detection Method in Complex Road Scenes[J].,2025,(10):30.[doi:10.20165/j.cnki.ISSN1673-629X.2025.0113]

更新日期/Last Update: 2025-10-10