[1]沈朕宇,朱凤华,王知学,等.基于高效特征提取和大感受野的无人机航拍图像目标检测[J].智能系统学报,2025,20(4):813-821.[doi:10.11992/tis.202405001]
SHEN Zhenyu,ZHU Fenghua,WANG Zhixue,et al.Uav aerial image target detection based on high-efficiency feature extraction and large receptive field[J].CAAI Transactions on Intelligent Systems,2025,20(4):813-821.[doi:10.11992/tis.202405001]
点击复制
SHEN Zhenyu,ZHU Fenghua,WANG Zhixue,et al.Uav aerial image target detection based on high-efficiency feature extraction and large receptive field[J].CAAI Transactions on Intelligent Systems,2025,20(4):813-821.[doi:10.11992/tis.202405001]
基于高效特征提取和大感受野的无人机航拍图像目标检测
《智能系统学报》[ISSN 1673-4785/CN 23-1538/TP] 卷:
20
期数:
2025年第4期
页码:
813-821
栏目:
学术论文—机器学习
出版日期:
2025-08-05
- Title:
- Uav aerial image target detection based on high-efficiency feature extraction and large receptive field
- Keywords:
- drone aerial images; small target detection; feature extraction; multi-scale variation; YOLOv8; context information; receptive field; loss function
- 分类号:
- TP391.4
- 文献标志码:
- 2025-2-26
- 摘要:
- 针对无人机航拍图像中存在小目标、目标遮挡、背景复杂的问题，提出一种基于高效特征提取和大感受野的目标检测网络(efficient feature and large receptive field network, EFLF-Net)。通过优化检测层架构降低小目标漏检率；在主干网络融合新的构建模块以提升特征提取效率；引入内容感知特征重组模块和大型选择性核网络，增强颈部网络对遮挡目标的上下文感知能力；采用Wise-IoU损失函数优化边界框回归稳定性。在VisDrone2019数据集上的实验结果表明，EFLF-Net较基准模型在平均精度上提高了5.2%。与已有代表性的目标检测算法相比，该方法对存在小目标、目标相互遮挡和复杂背景的无人机航拍图像有更好的检测效果。
- Abstract:
- Aiming at the problems of small targets, target occlusion and complex background in UAV aerial images, a target detection network based on high-efficiency feature extraction and large receptive field (EFLF-Net) was proposed. Firstly, the missed detection rate of small targets was reduced by optimizing the detection layer architecture. Then, the new building blocks were integrated in the backbone network to improve the efficiency of feature extraction. Then, a content-aware feature recombination module and a large selective kernel network were introduced to enhance the context-aware ability of the neck network for occluded targets. Finally, the Wise-IoU loss function was used to optimize the bounding box regression stability. Experimental results on the VisDrone2019 dataset show that EFLF-Net improves the average precision by 5.2% compared with the basic algorithm. Compared with the existing representative target detection algorithms, the proposed method has better detection effects for UAV aerial images with small targets, mutual occlusion of targets and complex backgrounds.
- 参考文献/References:
-
[1] 何宇豪, 易明发, 周先存, 等. 基于改进的Yolov5的无人机图像小目标检测[J]. 智能系统学报, 2024, 19(3): 635-645.
HE Yuhao, YI Mingfa, ZHOU Xiancun, et al. UAV image small-target detection based on improved Yolov5[J]. CAAI transactions on intelligent systems, 2024, 19(3): 635-645.
[2] 刘威, 靳宝, 周璇, 等. 基于特征融合及自适应模型更新的相关滤波目标跟踪算法[J]. 智能系统学报, 2020, 15(4): 714-721.
LIU Wei, JIN Bao, ZHOU Xuan, et al. Correlation filter target tracking algorithm based on feature fusion and adaptive model updating[J]. CAAI transactions on intelligent systems, 2020, 15(4): 714-721.
[3] 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. Las Vegas: IEEE, 2016: 779-788.
[4] TAN Ling, WU Hui, XU Zifeng, et al. Multi-object garbage image detection algorithm based on SP-SSD[J]. Expert systems with applications, 2025, 263: 125773.
[5] LIN T Y, GOYAL P, GIRSHICK R, et al. Focal loss for dense object detection[J]. IEEE transactions on pattern analysis and machine intelligence, 2020, 42(2): 318-327.
[6] CUI Jian, ZHANG Xinle, ZHANG Jiahuan, et al. Weed identification in soybean seedling stage based on UAV images and Faster R-CNN[J]. Computers and electronics in agriculture, 2024, 227: 109533.
[7] KARAKO K, MIHARA Y, ARITA J, et al. Automated liver tumor detection in abdominal ultrasonography with a modified faster region-based convolutional neural networks (Faster R-CNN) architecture[J]. Hepatobiliary surgery and nutrition, 2022, 11(5): 675-683.
[8] 秦振, 李学伟, 刘宏哲. 基于改进SSD的鲁棒小目标检测算法[J]. 东北师大学报(自然科学版), 2023, 55(4): 59-66.
QIN Zhen, LI Xuewei, LIU Hongzhe. Robust small tar-get detection algorithm based on improved SSD[J]. Journal of Northeast Normal University(natural science edition), 2023, 55(4): 59-66.
[9] LEE S S, LIM L G, PALAIAHNAKOTE S, et al. Oil palm tree detection in UAV imagery using an enhanced RetinaNet[J]. Computers and electronics in agriculture, 2024, 227: 109530.
[10] 邓姗姗, 黄慧, 马燕. 基于改进Faster R-CNN的小目标检测算法[J]. 计算机工程与科学, 2023, 45(5): 869-877.
DENG Shanshan, HUANG Hui, MA Yan. A small object detection algorithm based on improved Faster R-CNN[J]. Computer engineering & science, 2023, 45(5): 869-877.
[11] 吴明杰, 云利军, 陈载清, 等. 改进YOLOv5s的无人机视角下小目标检测算法[J]. 计算机工程与应用, 2019, 60(2): 191-199.
WU Mingjie, YUN Lijun, CHEN Zaiqing, et al. Improved YOLOv5s small target detection algorithm in UAV view[J]. Computer engineering and applications, 2019, 60(2): 191-199.
[12] WANG Xin, HE Ning, HONG Chen, et al. Improved YOLOX-X based UAV aerial photography object detection algorithm[J]. Image and vision computing, 2023, 135: 104697.
[13] 牛为华, 魏雅丽. 基于改进YOLOv 7的航拍小目标检测算法[J]. 电光与控制, 2024, 31(1): 117-122.
NIU Weihua, WEI Yali. Small target detection in aerial photography images based on improved YOLOv7 algorithm[J]. Electronics optics & control, 2024, 31(1): 117-122.
[14] TERVEN J, CóRDOVA-ESPARZA D M, ROMERO-GONZáLEZ J A. A comprehensive review of YOLO architectures in computer vision: from YOLOv1 to YOLOv8 and YOLO-NAS[J]. Machine learning and knowledge extraction, 2023, 5(4): 1680-1716.
[15] KRIEGEL J, DEJAM J, DURTH H, et al. Zur strafbarkeit von datenfunden im darknet[J]. Datenschutz und datensicherheit-DuD, 2024, 48(12): 769-774.
[16] SHEN Kenan, ZHAO Dongbiao. Fault analysis and fault degree evaluation via an improved ResNet method for aircraft hydraulic system[J]. Scientific reports, 2025, 15: 4132.
[17] FENG Dapeng, ZHUANG Xuebin, CHEN Zhiqiang, et al. Position information encoding FPN for small object detection in aerial images[J]. Neural computing and applications, 2024, 36(26): 16023-16035.
[18] LIU Shu, QI Lu, QIN Haifang, 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.
[19] 伍麟, 郝鸿宇, 宋友. 基于计算机视觉的工业金属表面缺陷检测综述[J]. 自动化学报, 2024, 50(7): 1261-1283.
WU Lin, HAO Hongyu, SONG You. A review of industrial metal surface defect detection based on computer vision [J]. IEEE/CAA journal of automatica sinica, 2019, 50(7): 1261-1283.
[20] SUNKARA R, LUO Tie. No more strided convolutions or pooling: a new CNN building block for low-resolution images and small objects[EB/OL]. (2022-08-07)[2024-04-01]. https://arxiv.org/abs/2208.03641.
[21] DING Zhipeng, WANG Ben, SUN Shuifa, et al. Improved SmapGAN remote sensing image map generation based on multi-head self-attention and carafe[J]. Journal of applied remote sensing, 2024, 18(1): 014526.
[22] HOU Yue, ZHANG Zhihao, DU Lixia, et al. A fully locally selective large kernel network for traffic video detection[J]. Measurement, 2025, 242: 115779.
[23] WANG Chenghao, LUO Zhongqiang, QI Ziyuan. Transformer oil leakage detection with sampling-WIoU module[J]. The journal of supercomputing, 2024, 80(6): 7349-7368.
[24] HUANG Zixin, TAO Xuesong, LIU Xinyuan. NAN-DETR: noising multi-anchor makes DETR better for object detection[J]. Frontiers in neurorobotics, 2024, 18: 1484088.
[25] MARAPATLA A D K, ILAVARASAN E. An effective attack detection framework using multi-scale depth-wise separable 1DCNN via fused grasshopper-based lemur optimizer in IoT routing system[J]. Intelligent decision technologies, 18(3): 1741-1762.
[26] WANG Xin, HE Ning, HONG Chen, et al. YOLO-ERF: lightweight object detector for UAV aerial images[J]. Multimedia systems, 2023, 29(6): 3329-3339.
[27] ZHU Xingfei, WANG Qimeng, ZHANG Bufan, et al. An improved feature enhancement CenterNet model for small object defect detection on metal surfaces[J]. Advanced theory and simulations, 2024, 7(8): 2301230.
[28] NAWAZ M, NAZIR T, MASOOD M, et al. Analysis of brain MRI images using improved CornerNet approach[J]. Diagnostics, 2021, 11(10): 1856.
[29] WANG Zhaodi, YANG Shuqiang, QIN Huafeng, et al. CCW-YOLO: a modified YOLOv5s network for pedestrian detection in complex traffic scenes[J]. Information, 2024, 15(12): 762.
[30] ZHANG Hongtao, ZHENG Li, TAN Lian, et al. YOLOX-S-TKECB: a Holstein cow identification detection algorithm[J]. Agriculture, 2024, 14(11): 1982.
- 相似文献/References:
-
[1]赵文清,周震东,翟永杰.基于反卷积和特征融合的SSD小目标检测算法[J].智能系统学报,2020,15(2):310.[doi:10.11992/tis.201905035]
ZHAO Wenqing,ZHOU Zhendong,ZHAI Yongjie.SSD small target detection algorithm based on deconvolution and feature fusion[J].CAAI Transactions on Intelligent Systems,2020,15():310.[doi:10.11992/tis.201905035]
[2]赵文清,孔子旭,赵振兵.隔级融合特征金字塔与CornerNet相结合的小目标检测[J].智能系统学报,2021,16(1):108.[doi:10.11992/tis.202004033]
ZHAO Wenqing,KONG Zixu,ZHAO Zhenbing.Small target detection based on a combination of feature pyramid and CornerNet[J].CAAI Transactions on Intelligent Systems,2021,16():108.[doi:10.11992/tis.202004033]
[3]毛莺池,唐江红,王静,等.基于Faster R-CNN的多任务增强裂缝图像检测方法[J].智能系统学报,2021,16(2):286.[doi:10.11992/tis.201910004]
MAO Yingchi,TANG Jianghong,WANG Jing,et al.Multi-task enhanced dam crack image detection based on Faster R-CNN[J].CAAI Transactions on Intelligent Systems,2021,16():286.[doi:10.11992/tis.201910004]
[4]齐鹏宇,王洪元,张继,等.基于改进FCOS的拥挤行人检测算法[J].智能系统学报,2021,16(4):811.[doi:10.11992/tis.202010012]
QI Pengyu,WANG Hongyuan,ZHANG Ji,et al.Crowded pedestrian detection algorithm based on improved FCOS[J].CAAI Transactions on Intelligent Systems,2021,16():811.[doi:10.11992/tis.202010012]
[5]李海丰,李纪霖,王怀超,等.复杂机场道面外来异物高精度实时检测算法[J].智能系统学报,2023,18(3):525.[doi:10.11992/tis.202110014]
LI Haifeng,LI Jilin,WANG Huaichao,et al.High-precision real-time detection algorithm for foreign object debris on complex airport pavements[J].CAAI Transactions on Intelligent Systems,2023,18():525.[doi:10.11992/tis.202110014]
[6]何宇豪,易明发,周先存,等.基于改进的Yolov5的无人机图像小目标检测[J].智能系统学报,2024,19(3):635.[doi:10.11992/tis.202210032]
HE Yuhao,YI Mingfa,ZHOU Xiancun,et al.UAV image small-target detection based on improved Yolov5[J].CAAI Transactions on Intelligent Systems,2024,19():635.[doi:10.11992/tis.202210032]
备注/Memo
收稿日期:2024-5-3。
基金项目:国家自然科学基金项目(U24A20277); 北京市自然科学基项目(L241016); 重庆市交通科技项目(CQJT-CZKJ2024-04).
作者简介:沈朕宇，硕士研究生，主要研究方向为图像处理与目标检测。E-mail：2216825930@qq.com。;朱凤华，副研究员，博士，主要研究方向为智能交通、云计算与大数据分析。E-mail：fenghua.zhu@ia.ac.cn。;熊刚，研究员、博士生导师，主要研究方向为人工智能、智能控制与管理。获吴文俊人工智能奖、中国自动化学会科技奖等10余项。发表学术论文450余篇，出版专著共3部，授权PCT 6项，授权专利90余项，登记软著90余项。E-mail：gang.xiong@ia.ac.cn。
通讯作者:熊刚. E-mail：gang.xiong@ia.ac.cn
更新日期/Last Update:
1900-01-01