遥感影像目标检测的轻量化密集连接网络

doi:10.13474/j.cnki.11-2246.2022.0291

测绘通报 ›› 2022, Vol. 0 ›› Issue (10): 37-43,55.doi: 10.13474/j.cnki.11-2246.2022.0291

遥感影像目标检测的轻量化密集连接网络

刘继^1,2,3, 杨军^1,2,3

1. 兰州交通大学测绘与地理信息学院, 甘肃兰州 730070;
2. 地理国情监测技术应用国家地方联合工程研究中心, 甘肃兰州 730070;
3. 甘肃省地理国情监测工程实验室, 甘肃兰州 730070

收稿日期:2021-11-09 发布日期:2022-11-02
通讯作者: 杨军。E-mail:yangj@mail.lzjtu.cn
作者简介:刘继(1996-),男,硕士生,主要研究方向为计算机视觉、遥感影像分析与检测等。E-mail:877522840@qq.com
基金资助:
国家自然科学基金(61862039);甘肃省科技计划(20JR5RA429);2021年度中央引导地方科技发展资金(2021-51);兰州市人才创新创业项目(2020-RC-22);兰州交通大学天佑创新团队(TY202002)

A lightweight dense connection network for object detection of remote sensing images

LIU Ji^1,2,3, YANG Jun^1,2,3

1. Faculty of Geomatics, Lanzhou Jiaotong University, Lanzhou 730070, China;
2. National-Local Joint Engineering Research Center of Technologies and Applications for National Geographic State Monitoring, Lanzhou 730070, China;
3. Gansu Provincial Engineering Laboratory for National Geographic State Monitoring, Lanzhou 730070, China

Received:2021-11-09 Published:2022-11-02

摘要/Abstract

摘要： 针对无人机遥感影像中可能出现的多个密集的小目标,在检测时会出现误检、漏检的难点问题,本文提出了一种基于YOLO v4的具有密集连接网络模块的遥感影像轻量化目标检测算法,实现了对无人机遥感影像中车辆小目标的高精度识别。首先,对YOLO v4主干网络CSP Darknet53的卷积层采用密集连接、稀疏连接两种处理方式,加强特征的提取和重复使用,以缓解梯度消失问题;然后,对此模块进行模型剪裁,减少网络层数并定义为新的密集连接网络模块;最后,在NWPU-VHR-10数据集和笔者所在课题组制作的Vehicle-850无人机影像数据集上进行了对比试验并取得了较好的效果。本文改进后的网络结构在提高遥感影像目标检测准确率的同时,缩短了网络模型的收敛时间,减少了模型占用的内存空间,提高了遥感影像中目标检测的速度。

关键词: 目标检测, 无人机遥感影像, YOLO v4, 轻量化, 密集连接

Abstract: Two issues arising out of the small objects, detected within UAV remote sensing images, sometimes happening to be multiple and dense increases in both the misdetection and miss out rates. A lightweight target detection algorithm is proposed based on YOLO v4 with a dense connection network module to realize high precision identification of the small targets. Firstly, feature reuse and feature extraction enhancement are facilitated by adjusting the convolutional layers of CSP Darknet53 (backbone network of YOLO v4) to work in two modes of actions:dense connection and sparse connection; the gradient disappearance problem is also alleviated. Secondly, the proposed module tailors the models to fit into a reduced number of network layers; then the resulting module is redefined as the new densely connected network module. Experiments are carried out on the NWPU-VHR-10 and Vihicle-850 UAV image datasets, a courtesy of our research group. Our model has the upper hand in terms of the number of the accurately detected small targets within the remote sensing images; our implementation is also cost-effective in terms of the network model convergence time and memory consumption. A significant improvement in the speed of detection is gained.

Key words: object detection, unmanned aerial vehicle remote sensing images, YOLO v4, lightweight, dense connection

中图分类号:

P237

刘继, 杨军. 遥感影像目标检测的轻量化密集连接网络[J]. 测绘通报, 2022, 0(10): 37-43,55.

LIU Ji, YANG Jun. A lightweight dense connection network for object detection of remote sensing images[J]. Bulletin of Surveying and Mapping, 2022, 0(10): 37-43,55.

参考文献

[1] LI Ke,WAN Gang,CHENG Gong,et al.Object detection in optical remote sensing images:a survey and a new benchmark[J].ISPRS Journal of Photogrammetry and Remote Sensing,2020,159:296-307.
[2] ZHANG Zhongyu,LIU Yunpeng,LIU Tianci,et al.DAGN:a real-time UAV remote sensing image vehicle detection framework[J].IEEE Geoscience and Remote Sensing Letters,2020,17(11):1884-1888.
[3] JIANG Zicong,ZHAO Liquan,LI Shuaiyang,et al.Real-time object detection method based on improved YOLOv4-tiny[J].Computer Vision and Pattern Recognition,2020(1):04244.
[4] DALAL N,TRIGGS B.Histograms of oriented gradients for human detection[C]//Proceedings of 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition.San Diego,CA,USA:IEEE,2005:886-893.
[5] LOWE D G.Distinctive image features from scale-invariant keypoints[J].International Journal of Computer Vision,2004,60(2):91-110.
[6] 慎利,唐宏,王世东,等.结合空间像素模板和Adaboost算法的高分辨率遥感影像河流提取[J].测绘学报,2013,42(3):344-350.
[7] BOYKOV Y,HUTTENLOCHER D.A Bayesian framework for model based tracking[C]//Proceedings of 2000 IEEE Conference on Computer Vision and Pattern Recognition.Hilton Head,SC:IEEE Computer Society,2000:697-704.
[8] LIU Wei,ANGUELOV D,ERHAN D,et al.SSD:single shot MultiBox detector[M]//Computer Vision.Cham:Springer,2016:21-37.
[9] REDMON J,DIVVALA S,GIRSHICK R,et al.You only look once:unified,real-time object detection[C]//Proceedings of 2016 IEEE Conference on Computer Vision and Pattern Recognition.Las Vegas,NV,USA:IEEE,2016:779-788.
[10] GIRSHICK R,DONAHUE J,DARRELL T,et al.Rich feature hierarchies for accurate object detection and semantic segmentation[C]//Proceedings of 2014 IEEE Conference on Computer Vision and Pattern Recognition.Columbus,OH,USA:IEEE,2014:580-587.
[11] REN Shaoqing,HE Kaiming,GIRSHICK R,et al.Faster R-CNN:towards real-time object detection with region proposal networks[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2017,39(6):1137-1149.
[12] DAI J F,LI Y,HE K M,et al.R-FCN:object detection via region-based fully convolutional networks[C]//Proceedings of 2016 IEEE Conference on Computer Vision and Pattern Recognition,Las Vegas,NV:IEEE Computer Society,2016:379-387.
[13] REDMON J,FARHADI A.YOLO9000:better,faster,stronger[C]//Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition.Honolulu,HI,USA:IEEE,2017:6517-6525.
[14] REDMON J,FARHADI A.YOLO v3:an incremental improvement[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2018,15:1125-1131.
[15] BOCHKOVSKIY A,WANG C Y,LIAO H Y M.YOLOv4:optimal speed and accuracy of object detection[J].Computer Vision and Pattern Recognition,2020:10934.
[16] HUANG Gao,LIU Zhuang,VAN DER MAATEN L,et al.Densely connected convolutional networks[C]//Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition.Honolulu,HI,USA:IEEE,2017:2261-2269.
[17] HE Kaiming,ZHANG Xiangyu,REN Shaoqing,et al.Deep residual learning for image recognition[C]//Proceedings of 2016 IEEE Conference on Computer Vision and Pattern Recognition,Las Vegas,NV,USA:IEEE,2016:770-778.
[18] CHENG Gong,HAN Junwei,ZHOU Peicheng,et al.Multi-class geospatial object detection and geographic image classification based on collection of part detectors[J].ISPRS Journal of Photogrammetry and Remote Sensing,2014,98:119-132.
[19] DUAN Kaiwen,BAI Song,XIE Lingxi,et al.CenterNet:keypoint triplets for object detection[C]//Proceedings of 2019 IEEE/CVF International Conference on Computer Vision.Seoul,Korea (South):IEEE,2019:6568-6577.
[20] ZHANG Shaojing,MU Xiaodong,KOU Guangjie,et al.Object detection based on efficient multiscale auto-inference in remote sensing images[J].IEEE Geoscience and Remote Sensing Letters,2021,18(9):1650-1654.

遥感影像目标检测的轻量化密集连接网络

A lightweight dense connection network for object detection of remote sensing images

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

本文评价

[1]	张光伟, 吴昊, 郭震冬. 实景三维多源数据场景融合[J]. 测绘通报, 2022, 0(8): 155-159.
[2]	柳林, 孙毅, 李万武. 基于CNN海上钻井平台检测模型的构建及训练算法分析[J]. 测绘通报, 2022, 0(7): 26-32,99.
[3]	文奴, 郭仁忠, 贺彪, 万远. YOLO v4框架下Multi-Patch多帧增量式交通视频目标检测[J]. 测绘通报, 2022, 0(5): 38-44.
[4]	潘建平, 李鑫, 孙博文, 胡勇, 李明明. 基于注意力密集连接金字塔网络的新增建设用地变化检测[J]. 测绘通报, 2022, 0(3): 41-46,59.
[5]	段晨阳, 冯建中, 全斌, 白林燕, 王盼盼. 利用深度学习进行GF-6影像枣园检测识别[J]. 测绘通报, 2022, 0(3): 54-59.
[6]	程瑞, 高建, 邢强, 孙中昶. 一种优化的Faster R-CNN小目标检测方法[J]. 测绘通报, 2021, 0(9): 21-27.
[7]	梁哲恒, 邓鹏, 姜福泉, 盛森, 魏汝兰, 谢刚生. 基于卷积神经网络的无人机影像违章建筑检测应用[J]. 测绘通报, 2021, 0(4): 111-115.
[8]	马京振, 陈换新, 朱新铭, 张付兵. 利用Faster R-CNN进行立交桥自动识别与定位[J]. 测绘通报, 2021, 0(3): 28-32,86.
[9]	侯翘楚, 李必军, 蔡毅. 高分辨率遥感影像的车道级高精地图要素提取[J]. 测绘通报, 2021, 0(3): 38-43.
[10]	何志伟, 唐伯惠, 王涛, 王晓红, 于伯华, 李闯, 邓仕雄. 剔除无人机影像BRISK特征误匹配点对算法[J]. 测绘通报, 2021, 0(10): 78-82.
[11]	王文凯, 黄昆学, 李梦, 王业明. 倾斜摄影测量数据轻量化处理技术研究[J]. 测绘通报, 2020, 0(9): 123-126.
[12]	王玮. 基于三维GIS的铁路BIM空间信息系统构建及其工程应用[J]. 测绘通报, 2020, 0(7): 138-142.
[13]	张永宏, 蔡朋艳, 陶润喆, 王剑庚, 田伟. 基于改进U-Net网络的遥感图像云检测[J]. 测绘通报, 2020, 0(3): 17-20,34.
[14]	沈雨, 仇宇轩, 于正浩. 分部加权的行道树检测网络构建[J]. 测绘通报, 2020, 0(2): 77-81,101.
[15]	杨梦圆, 刘伟, 尹鹏程, 谢梦. 深度卷积网络支持下的遥感影像井盖部件检测[J]. 测绘通报, 2019, 0(8): 78-81,87.