[1]马田源,孙 涵.改进 RetinaNet 特征融合方式的无人机检测方法[J].计算机技术与发展,2022,32(12):103-109.[doi:10. 3969 / j. issn. 1673-629X. 2022. 12. 016]
　MA Tian-yuan,SUN Han.An Improved Feature Fusion Method for Drone Detection Based on RetinaNet Extraction[J].,2022,32(12):103-109.[doi:10. 3969 / j. issn. 1673-629X. 2022. 12. 016]
点击复制

改进 RetinaNet 特征融合方式的无人机检测方法()
分享到：

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

卷:
32
期数:
2022年12期
页码:
103-109
栏目:
网络空间安全
出版日期:
2022-12-10

文章信息/Info

Title:
An Improved Feature Fusion Method for Drone Detection Based on RetinaNet Extraction
文章编号:
1673-629X(2022)12-0103-07
作者:
马田源; 孙 涵
南京航空航天大学 计算机科学与技术学院,江苏 南京 211106
Author(s):
MA Tian-yuan; SUN Han
School of Computer Science and Technology,Nanjing University of Aeronautics and Astronautics,Nanjing 211106,China
关键词:
小目标检测; 无人机检测; RetinaNet; 特征融合; 深度学习
Keywords:
small object detection; drone detection; RetinaNet; feature fusion; deep learning103
分类号:
TP301
DOI:
10. 3969 / j. issn. 1673-629X. 2022. 12. 016
摘要:
常见的目标检测方法如 R-CNN 系列方法、YOLO 系列方法和 RetinaNet 等,虽然在通用数据集上有着不俗表现,但是在无人机小目标检测任务中的表现却不尽如人意。 分析原因是这些方法采用的传统特征金字塔融合网络 FPN 存在着上采样失真、语义信息衰减以及深层语义差异的问题,导致目标检测网络未能获取足够有辨识度的特征,致使其在无人机小目标检测任务中表现不佳。 对此,该文提出了一种基于 RetinaNet 网络的多尺度特征融合方法。 该方法采用像素洗牌上采样模块构建了像素洗牌融合网络,并且引入了深层语义增强模块,可在多尺度特征融合阶段提升无人机小目标在网络浅层的特征表示效果,进而提升深度神经网络对无人机小目标的检测性能。 最后在自建蜂群无人机数据集上的实验结果显示,引入新的特征融合方法之后,网络对无人机的检测精度达到 91. 2% ,提升了 1. 7% 。
Abstract:
Although most object detection methods like R-CNN,YOLO and RetinaNet have outstanding performance in general datasets,but their performance in small drone detection task is not satisfactory. The reason is that the traditional feature pyramid fusion networkFPN adopted by these methods has problems of up-sampling distortion,semantic information attenuation and deep semantic difference,which leads to the failure of object detection network to acquire enough identifiable features,resulting in its poor performance in smalldrone detection task. Therefore,we propose a multi-scale feature fusion method based on RetinaNet network. In this method,the up-sampling module of pixel shuffle unit is used to construct the pixel shuffle feature fusion network, and the high - level semanticenhancement module is introduced to improve the feature representation effect of small drone in the shallow layer of the network in thestage of multi - scale feature fusion. Then the deep neural network can improve the detection performance of small drone. Finally,experiments on the self-built drone dataset show that the network detection accuracy of drone can reach 91. 2% ,which increases by 1. 7% .

相似文献/References:

[1]陶晓力,刘宁钟,沈家全.基于深度信息融合的航拍车辆检测[J].计算机技术与发展,2019,29(09):117.[doi:10. 3969 / j. issn. 1673-629X. 2019. 09. 023]
　TAO Xiao-li,LIU Ning-zhong,SHEN Jia-quan.Aerial Vehicle Detection Based on Depth Information Fusion[J].,2019,29(12):117.[doi:10. 3969 / j. issn. 1673-629X. 2019. 09. 023]
[2]产世兵,刘宁钟,沈家全.基于轻量级网络的 PCB 元器件检测[J].计算机技术与发展,2020,30(10):14.[doi:10. 3969 / j. issn. 1673-629X. 2020. 10. 003]
　CHAN Shi-bing,LIU Ning-zhong,SHEN Jia-quan.PCB Component Detection Based on Lightweight Network[J].,2020,30(12):14.[doi:10. 3969 / j. issn. 1673-629X. 2020. 10. 003]
[3]杨朝晨,陈佳悦,邢 可,等.基于改进的 DSSD 的小目标检测算法研究[J].计算机技术与发展,2022,32(06):63.[doi:10. 3969 / j. issn. 1673-629X. 2022. 06. 011]
　YANG Zhao-chen,CHEN Jia-yue,XING Ke,et al.Small Target Detection Algorithm Based on Improved DSSD[J].,2022,32(12):63.[doi:10. 3969 / j. issn. 1673-629X. 2022. 06. 011]
[4]刘梦菲,毛建华,陆小锋.基于注意力和上下文感知的海面渔桩检测[J].计算机技术与发展,2023,33(08):144.[doi:10. 3969 / j. issn. 1673-629X. 2023. 08. 021]
　LIU Meng-fei,MAO Jian-hua,LU Xiao-feng.Detection of Fishing Piles on Sea Surface Based on Attention and Context Awareness[J].,2023,33(12):144.[doi:10. 3969 / j. issn. 1673-629X. 2023. 08. 021]
[5]文 青,伍 欣,敖 斌,等.基于航空图像的目标检测算法 Trans_YOLOv5[J].计算机技术与发展,2024,34(01):77.[doi:10. 3969 / j. issn. 1673-629X. 2024. 01. 012]
　WEN Qing,WU Xin,AO Bin,et al.Target Detection Algorithm Trans_YOLOv5 Based on Aerial Image[J].,2024,34(12):77.[doi:10. 3969 / j. issn. 1673-629X. 2024. 01. 012]
[6]魏雅丽,牛为华.改进 YOLOv5s 的轻量化航拍小目标检测算法[J].计算机技术与发展,2024,34(02):53.[doi:10. 3969 / j. issn. 1673-629X. 2024. 02. 008]
　WEI Ya-li,NIU Wei-hua.Aerial Small Target Detection Based on Improved YOLOv5s Lightweight Algorithm[J].,2024,34(12):53.[doi:10. 3969 / j. issn. 1673-629X. 2024. 02. 008]
[7]赵 侃,汪慧兰,郭娇娇,等.基于 DTA-FSAF 的无人机小目标检测研究[J].计算机技术与发展,2024,34(04):101.[doi:10. 3969 / j. issn. 1673-629X. 2024. 04. 016]
　ZHAO Kan,WANG Hui-lan,GUO Jiao-jiao,et al.Research on Small Object Detection of UAV Based on DTA-FSAF[J].,2024,34(12):101.[doi:10. 3969 / j. issn. 1673-629X. 2024. 04. 016]
[8]耿 文,孙 涵.一种 Anchor Free 的无人机检测方法[J].计算机技术与发展,2021,31(01):54.[doi:10. 3969 / j. issn. 1673-629X. 2021. 01. 010]
　GENG Wen,SUN Han.An Anchor Free Method for Detecting Drones[J].,2021,31(12):54.[doi:10. 3969 / j. issn. 1673-629X. 2021. 01. 010]
[9]杨 健,孙 涵.基于深度特征提取的无人机检测算法[J].计算机技术与发展,2021,31(11):71.[doi:10. 3969 / j. issn. 1673-629X. 2021. 11. 012]
　YANG Jian,SUN Han.Drone Detection Based on Depth Feature Extraction[J].,2021,31(12):71.[doi:10. 3969 / j. issn. 1673-629X. 2021. 11. 012]

更新日期/Last Update: 2022-12-10