| 1,675 | 0 | 103 |
| 下载次数 | 被引频次 | 阅读次数 |
针对无人机航拍图像小目标检测中面临的目标尺寸微小、分布密集、特征信息模糊以及复杂背景干扰等挑战，对YOLO11模型进行改进。改进主要聚焦在3个方面：设计了渐进式多尺度特征提取模块CSMAConv(Channel Split Multi-scale Aggregation Convolution),对原有C3k2模块进行改造，通过通道分割与多尺度卷积级联结构，既保留了原始特征信息又扩大了感受野范围，显著增强了对多尺度细节特征的捕获能力；提出了新的特征金字塔结构HEFFPN(Hierarchical Efficient Fusion Feature Pyramid Network),通过构建额外的跨尺度特征融合路径并引入SBA(Selective Boundary Aggregation)特征融合模块，频繁整合不同尺度信息，提升了特征融合能力；设计了共享增强型检测头(Shared Enhanced Detection-Head, SED-Head),在降低计算复杂度的同时提高了小目标检测效率。在VisDrone-DET2019数据集的实验中，改进后的模型较YOLO11s在参数量降低51.1%的同时，mAP@0.5和mAP@0.5:0.95分别提高了0.099和0.07,验证了模型的有效性。
Abstract：To address the challenges of tiny target size, dense distribution, blurred feature information, and complex background interference in small target detection of UAV aerial images, the YOLO11 model is improved. The enhancements focus on three key aspects: a progressive multi-scale feature extraction module called Channel Split Multi-scale Aggregation Convolution(CSMAConv) is designed to modify the original C3k2 module. By using channel splitting and a cascaded multi-scale convolution structure, the original feature information is preserved and the receptive field is expanded, which significantly enhances the ability to capture multi-scale detailed features. A new feature pyramid structure called Hierarchical Efficient Fusion Feature Pyramid Network(HEFFPN) is proposed. By constructing additional cross-scale feature fusion paths and introducing the Selective Boundary Aggregation(SBA) module, it more frequently integrates multi-scale information and improves the feature fusion capability. A Shared Enhanced Detection-Head(SED-Head) is designed to reduce computational complexity while improving the detection efficiency of small targets. Experiments on the VisDrone-DET2019 dataset demonstrate that, compared to YOLO11s, the improved model reduces the number of parameters by 51.1%, while increasing mAP@0.5 by 0.099 and mAP@0.5:0.95 by 0.07, validating the effectiveness of the proposed method.
[1] PENG C,ZHU M,REN H E,et al.Small Object Detection Method Based on Weighted Feature Fusion and CSMA Attention Module[J].Electronics,2022,11(16):2546.
[2] ZHANG Q,ZHANG H Y,LU X W.Adaptive Feature Fusion for Small Object Detection[J].Applied Sciences,2022,12(22):11854.
[3] XU H,ZHENG W L,LIU F X,et al.Unmanned Aerial Vehicle Perspective Small Target Recognition Algorithm Based on Improved YOLOv5[J].Remote Sensing,2023,15(14):3583.
[4] LI Y T,FAN Q S,HUANG H S,et al.A Modified YOLOv8 Detection Network for UAV Aerial Image Recognition[J].Drones,2023,7(5):304.
[5] HAN K,WANG Y H,TIAN Q,et al.GhostNet:More Features From Cheap Operations[C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).Seattle:IEEE,2020:1577-1586.
[6] WU S X,LU X Y,GUO C C,et al.Accurate UAV Small Object Detection Based on HRFPN and EfficentVMamba[J].Sensors,2024,24(15):4966.
[7] GU A,DAO T.Mamba:Linear-time Sequence Modeling with Selective State Spaces[EB/OL].(2024-05-21) [2025-03-25].https://arxiv.org/abs/2312.00752.
[8] BU Y C,YE H R,TIE Z X,et al.OD-YOLO:Robust Small Object Detection Model in Remote Sensing Image with a Novel Multi-scale Feature Fusion[J].Sensors,2024,24(11):3596.
[9] SHI Y L,JIA Y,ZHANG X H.FocusDet:An Efficient Object Detector for Small Object[J].Scientific Reports,2024,14(1):10697.
[10] LI Y D,YAN H,LI D,et al.Robust Miner Detection in Challenging Underground Environments:An Improved YOLOv11 Approach[J].Applied Sciences,2024,14(24):11700.
[11] KHANAM R,HUSSAIN M.YOLOv11:An Overview of the Key Architectural Enhancements[EB/OL].(2024-10-23) [2025-03-25].https://www.arxiv.org/abs/2410.17725.
[12] CHOLLET F.Xception:Deep Learning with Depthwise Separable Convolutions[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).Honolulu:IEEE,2017:1800-1807.
[13] TANG F L,XU Z X,HUANG Q M,et al.DuAT:Dual-aggregation Transformer Network for Medical Image Segmentation[C]// Pattern Recognition and Computer Vision.Xiamen:Springer,2024:343-356.
[14] CHEN J R,KAO S H,HE H,et al.Run,Don’t Walk:Chasing Higher FLOPS for Faster Neural Networks[C]//2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).Vancouver:IEEE,2023:12021-12031.
[15] 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 (CVPR).Honolulu:IEEE,2017:936-944.
[16] 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.Salt Lake City:IEEE,2018:8759-8768.
[17] LI Y G,LI Q,PAN J,et al.SOD-YOLO:Small-object-detection Algorithm Based on Improved YOLOv8 for UAV Images[J].Remote Sensing,2024,16(16):3057.
[18] 王雪秋，高焕兵，郏泽萌.改进YOLOv8的道路缺陷检测算法[J].计算机工程与应用，2024,60(17):179-190.
[19] CHEN Z X,HE Z W,LU Z M.DEA-Net:Single Image Dehazing Based on Detail-enhanced Convolution and Content-guided Attention[J].IEEE Transactions on Image Processing,2024,33:1002-1015.
[20] WU Y X,HE K M.Group Normalization[C]// Computer Vision - ECCV 2018.Munich:Springer,2018:3-19.
[21] DU D W,ZHU P F,WEN L Y,et al.VisDrone-DET2019:The Vision Meets Drone Object Detection in Image Challenge Results[C]//2019 IEEE/CVF International Conference on Computer Vision Workshop (ICCVW).Seoul:IEEE,2019:213-226.
[22] GIRSHICK R.Fast R-CNN[C]//2015 IEEE International Conference on Computer Vision (ICCV).Santiago:IEEE,2015:1440-1448.
[23] ZHAO Y,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 (CVPR).Seattle:IEEE,2024:16965-16974.
[24] 范博淦，王淑青，陈开元.基于改进YOLOv8的航拍无人机小目标检测模型[J/OL].计算机应用，1-11[2025-03-25].http://kns.cnki.net/kcms/detail/51.1307.tp.20241017.1040.004.html.
[25] 梁燕，何孝武，邵凯，等.改进YOLOv8的无人机航拍图像目标检测算法[J].计算机工程与应用，2025,61(1):121-130.
[26] 付锦燚，张自嘉，孙伟，等.改进YOLOv8的航拍图像小目标检测算法[J].计算机工程与应用，2024,60(6):100-109.
[27] SELVARAJU R R,COGSWELL M,DAS A,et al.Grad-CAM:Visual Explanations from Deep Networks via Gradient-based Localization[C]//2017 IEEE International Conference on Computer Vision (ICCV).Venice:IEEE,2017:618-626.
基本信息:
DOI：
中图分类号:V19;TP391.41
引用信息:
[1]贾星宇,李大鹏.基于YOLO11的无人机航拍图像小目标检测算法[J].无线电工程,2025,55(08):1560-1570.
基金信息:
国家自然科学基金(62371245)~~