PCA-YOLO11:全域复杂环境的轻量目标检测

苏向阳; 汪洋; 谭森起; 储华珍; 潘博

doi:10.12382/bgxb.2025.0584

您当前的位置：

首页 >

文章列表页 >

PCA-YOLO11:全域复杂环境的轻量目标检测

更新时间：2026-02-10

- PCA-YOLO11:全域复杂环境的轻量目标检测
- PCA-YOLO11:Lightweight Object Detection in Multi-altitude Complex Environments
- 兵工学报 2025年46卷第S2期页码：250584
- 作者机构：
  
  1. 中兵智能创新研究院有限公司, 北京 100072
  2. 群体协同与自主实验室, 北京 100072
- 作者简介：
  
  * 邮箱: tristantsq@gmail.com
- 基金信息：
  
  国家自然科学基金项目(52202512)
- DOI：10.12382/bgxb.2025.0584
  中图分类号：
- 收稿：2025-07-01，
  
  网络首发：2026-02-03，
  
  纸质出版：2025
- 稿件说明：
移动端阅览
苏向阳, 汪洋, 谭森起, 等. PCA-YOLO11:全域复杂环境的轻量目标检测[J]. 兵工学报, 2025,46(S2):250584.

Xiangyang SU, Yang WANG, Senqi TAN, et al. PCA-YOLO11:Lightweight Object Detection in Multi-altitude Complex Environments[J]. Acta Armamentarii, 2025, 46(S2): 250584.
苏向阳, 汪洋, 谭森起, 等. PCA-YOLO11:全域复杂环境的轻量目标检测[J]. 兵工学报, 2025,46(S2):250584. DOI： 10.12382/bgxb.2025.0584.

Xiangyang SU, Yang WANG, Senqi TAN, et al. PCA-YOLO11:Lightweight Object Detection in Multi-altitude Complex Environments[J]. Acta Armamentarii, 2025, 46(S2): 250584. DOI： 10.12382/bgxb.2025.0584.

摘要

面向复杂环境中无人机对目标的实时精准检测需求

提出一种全域复杂环境的轻量化检测模型PCA-YOLO11(Pinwheel-shaped Context Aggregation YOLO11)。构建涵盖多高度、多视角及复杂背景干扰的全域视角特种车辆数据集

包含1311张图像与2729个特种车辆目标;设计融合风车形卷积与全局上下文聚合的特征增强模块

通过局部细节特征感知强化小目标特征提取与抗遮挡能力

并基于长程依赖建模同步增强抗背景干扰鲁棒性;引入尺度感知的动态损失函数

优化多尺度目标的定位精度;采用约束训练导向的层级通道剪枝策略

实现精度可控的轻量化。实验结果表明:PCA-YOLO11在自制数据集上的mAP@0.5-0.95达79.17%

较基础模型提高3.01%;轻量化版本参数量降低14.4%

精度损失仅0.45%

推理速度达128.8FPS

满足实时精准检测需求。

Abstract

To address the requirements for the real-time accurate detection of targets by UAVs

this paper proposes pinwheel-shaped context aggregation (PCA)-YOLO11 (PCA-YOLO11) model

a lightweight target detection model in multi-altitude complex environments.A comprehensive special vehicle dataset spanning the multi-altitude views with complex background interference is constructed

comprising 1 311 images and 2 729 targets.A feature enhancement module pinwheel-shaped context aggregation convolution integrating pinwheel-shaped convolution and global context aggregation is designed

which enhances small-target feature extraction and occlusion resistance through local detail perception while enhancing anti-background-interference robustness via long-range dependency modeling.Subsequently

a scale-based dynamic loss function is introduced to optimize the multi-scale target localization accuracy.Finally

a constrained training-guided hierarchical channel pruning strategy is used to achieve the compression of accuracy-preserving model.Experimental results demonstrate that PCA-YOLO11 achieves 79.17% mAP@0.5-0.95 on the custom dataset

outperforming the baseline model by 3.01%.The lightweight version reduces parameters by 14.4% with accuracy degradation of merely 0.45% and reaches 128.8 FPS inference speed

meeting the real-time accurate detection requirements.

关键词

Keywords

references

黄红 , 苏菡 , 闵鹏 . 融合多尺度特征的无人机图像中小目标检测算法 [J ] . 计算机科学 , 2025 , 52 ( 增刊1 ): 517 - 521 .

HUANG H , SU H , MIN P . Small target detection algorithm in UAV images integrating multi-scale features [J ] . Computer Science , 2025 , 52 ( S1 ): 517 - 521 . (in Chinese)

吴一全 , 童康 . 基于深度学习的无人机航拍图像小目标检测研究进展 [J ] . 航空学报 , 2025 , 46 ( 3 ): 030848 .

WU Y Q , TONG K . Research advances on deep learning-based small object detection in UAV aerial images [J ] . Acta Aeronautica et Astronautica Sinica , 2025 , 46 ( 3 ): 030848 . (in Chinese)

宋晓茹 , 刘康 , 高嵩 , 等 . 基于深度学习的军事目标识别算法综述 [J ] . 科学技术与工程 , 2022 , 22 ( 22 ): 9466 - 9475 .

SONG X R , LIU K , GAO S , et al. Overview of military target recognition algorithms based on deep learning [J ] . Science Technology and Engineering , 2022 , 22 ( 22 ): 9466 - 9475 . (in Chinese)

GIRSHICK R . Fast R-CNN [C ] //Proceedings of the 2015 IEEE International Conference on Computer Vision.Santiago, Chile:IEEE , 2015 : 1440 - 1448 .

ALI M L , ZHANG Z . The YOLO framework:a comprehensive review of evolution,applications,and benchmarks in object detection [J ] . Computers , 2024 , 13 ( 12 ): 336 .

KHANAM R , HUSSAIN M . Yv11:an overview of the key architectural enhancements:arXiv:2410:17725 [R ] . Ithaca,NY , US : Cornell University , 2024 , 2410 :17725.

于博文 , 吕明 . 改进的YOLOv3算法及其在军事目标检测中的应用 [J ] . 兵工学报 , 2022 , 43 ( 2 ): 345 - 354 .

YU B W , LÜ M . Improved YOLOv3 algorithm and its application military target detection [J ] . Acta Armamentarii , 2022 , 43 ( 2 ): 345 - 354 . (in Chinese)

梅礼坤 , 陈智利 , 牛恒 . 基于改进YOLOv5的坦克目标检测算法 [J ] . 激光杂志 , 2023 , 44 ( 3 ): 1 - 7 .

MEI L K , CHEN Z L , NIU H . Tank target detection algorithm based on improved YOLOv5 [J ] . Laser Journal , 2023 , 44 ( 3 ): 1 - 7 . (in Chinese)

郑陆石 , 胡晓锋 , 于伟国 , 等 . 基于改进YOLOv7-tiny的坦克车辆检测方法 [J ] . 兵器装备工程学报 , 2023 , 44 ( 12 ): 285 - 292 .

ZHENG L S , HU X F , YU W G , et al. Tank vehicle detection method based on improved YOLOv7-tiny [J ] . Journal of Ordnance Equipment Engineering , 2023 , 44 ( 12 ): 285 - 292 . (in Chinese)

李治林 , 杜玉军 , 王牌 . 基于YOLOv4改进算法的坦克装甲车辆目标检测 [J ] . 电光与控制 , 2024 , 31 ( 6 ): 105 - 111 .

LI Z L , DU Y J , WANG P . Tank and armored vehicle target detection based on improved YOLOv4 algorithm [J ] . Electronics Optics & Control , 2024 , 31 ( 6 ): 105 - 111 . (in Chinese)

谢国波 , 吴陈锋 , 林志毅 . 遥感军事坦克轻量化检测的MSG-YOLOv7算法 [J ] . 现代电子技术 , 2024 , 47 ( 19 ): 47 - 54 .

XIE G B , WU C F , LIN Z Y . Lightweight detection algorithm MSG-YOLOv7 for special vehicles in remote sensing images [J ] . Modern Electronics Technique , 2024 , 47 ( 19 ): 47 - 54 . (in Chinese)

冯迎宾 , 郭枭尊 , 晏佳华 . 基于多尺度注意力机制的无人机小目标检测算法 [J ] . 兵工学报 , 2025 , 46 ( 1 ): 12 - 21 .

FENG Y B , GUO X Z , YAN J H . Small UAV target detection algorithm based on multi-scale attention mechanism [J ] . Acta Armamentarii , 2025 , 46 ( 1 ): 12 - 21 . (in Chinese)

徐图南 , 高昂 , 陈昱成 , 等 . 面向战场低空目标的轻量化识别方法 [J ] . 兵工学报 , 2025 , 46 ( 2 ): 1 - 11 .

XU T N , GAO A , CHEN Y C , et al. A lightweight recognition method for low altitude targets in the battlefield [J ] . Acta Armamentarii , 2025 , 46 ( 2 ): 1 - 11 . (in Chinese)

LIU H H , YU Y H , LIU S Z , et al. A military object detection model of UAV reconnaissance image and feature visualization [J ] . Applied Sciences , 2022 , 12 ( 23 ): 12236 .

YANG J N , LIU S L , WU J J , et al. Pinwheel-shaped convolution and scale-based dynamic loss for infrared small target detection [C ] //Proceedings of the AAAI Conference on Artificial Intelligence.Philadelphia,PA, US:AAAI , 2025 : 9202 - 9210 .

LIU Y , LI H F , HU C , et al. Learning to aggregate multi-scale context for instance segmentation in remote sensing images [J ] . IEEE Transactions on Neural Networks and Learning Systems , 2024 , 36 ( 1 ): 595 - 609 .

FANG G F , MA X Y , SONG M L , et al. Depraph:towards any structural pruning [C ] //Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.Vancouver, Canada:IEEE , 2023 : 16091 - 16101 .

ZHUANG T , ZHANG Z X , HUANG Y H , et al. Neuron-level structured pruning using polarization regularizer [J ] . Advances in Neural Information Processing Systems , 2020 , 33 : 9865 - 9877 .

LIU Z , LI J G , SHEN Z Q , et al. Learning efficient convolutional networks through network slimming [C ] //Proceedings of the IEEE International Conference on Computer Vision.Venice, Italy:IEEE , 2017 : 2736 - 2744 .

LIN T Y , MAIRE M , BELONGIE S , et al. Microsoft COCO:common objects in context [C ] // Computer Vision-ECCV 2014:13th European Conference.Zurich,Switzerland:Springer , 2014 : 740 - 755 .

SOHAN M , SAI R T , RAMI R C V . A review on YOLOv8 and its advancements [C ] //Proceedings of International Conference on Data Intelligence and Cognitive Informatics.Tirunelveli, India:Springer , 2024 : 529 - 545 .

TIAN Y J , YE Q X , DOERMANN D . YOLOv12:attention-centric real-time object detectors:arXiv:2502.12524 [R/OL ] . Ithaca,NY , US : Cornell University , 2025 (2025-02-18)[2025-06-15 ] . https://doi.org/10.48550/arXiv.2502.12524 https://doi.org/10.48550/arXiv.2502.12524 .

浏览量

下载量

CNKI被引量

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于改进海洋捕食者算法的无人机三维航迹规划

无人机数据链带外电磁干扰3阶互调效应预测

基于YOLOv10n的轻量化红外小目标检测模型

无人机末端机动规避中远距空空导弹策略

基于微运动片段的履带式车辆多路面-三维度工况构建方法