基于多尺度融合特征卷积神经网络的遥感图像飞机目标检测

doi:10.11947/j.AGCS.2019.20180398

测绘学报 ›› 2019, Vol. 48 ›› Issue (10): 1266-1274.doi: 10.11947/j.AGCS.2019.20180398

基于多尺度融合特征卷积神经网络的遥感图像飞机目标检测

姚群力^1,2, 胡显^1,2, 雷宏¹

1. 中国科学院电子学研究所航天微波遥感系统部, 北京 100190;
2. 中国科学院大学电子电气与通信工程学院, 北京 100049

收稿日期:2018-08-26 修回日期:2019-05-05 出版日期:2019-10-20 发布日期:2019-10-24
作者简介:姚群力(1993-),男,硕士,研究方向为机器学习。E-mail:yaoqunli15@mails.ucas.ac.cn
基金资助:
国家自然科学基金青年基金（61422113； 61601437）；国家重点研发计划（2017YFB0502700）

Aircraft detection in remote sensing imagery with multi-scale feature fusion convolutional neural networks

YAO Qunli^1,2, HU Xian^1,2, Lei Hong¹

1. Department of Space Microwave Remote Sensing Systems, Institute of Electronics, Chinese Academy of Sciences, Beijing 100190, China;
2. School of Electronics, Electrical and Communication Engineering, University of Chinese Academy of Sciences, Beijing 100049, China

Received:2018-08-26 Revised:2019-05-05 Online:2019-10-20 Published:2019-10-24
Supported by:
The National Natural Science Foundation of China (Nos. 61422113;61601437);The National Key Research and Development Program of China (No. 2017YFB0502700)

摘要/Abstract

摘要： 飞机检测在遥感图像解译中具有重要的研究意义。针对现有目标检测算法对于复杂场景区域或飞机密集区域的小尺度飞机目标检测精度较低的问题，本文提出了一种端到端的多尺度特征融合飞机目标检测框架MultDet。该方法基于SSD多尺度检测框架，采用轻量级基础网络提取多尺度特征信息；然后设计反卷积特征融合模块，通过跳跃连接将高层语义特征与低层细节特征进行特征融合，得到结构层次丰富的多尺度融合特征；最后设计了一系列不同纵横比的候选框以适应多尺度飞机目标检测。本文在光学遥感图像数据集UCAS-AOD上进行数据分析试验，结果表明，MultDet512在飞机数据集上取得了94.8%的平均检测精度（average precision，AP），在Titan Xp GPU上达到0.050 0 s/img的检测速度。本文所提飞机目标检测算法在包含多种复杂场景的遥感图像中，能够实现多尺度飞机目标的高精度稳健检测。

关键词: 遥感图像, 飞机检测, 特征融合, 多尺度特征

Abstract: Aircraft detection in remote sensing images (RSIs) is a meaningful task. There are many problems in current detection methods, such as low accuracy in complex background and dense aircraft area, especially for small-scale aircraft. To solve these problems, an end-to-end aircraft detection method named MultDet is proposed in this paper. Based on single shot multibox detector (SSD), a lightweight baseline Network is used to extract multi-scale features for its powerful ability in feature extraction. To obtain the feature maps with enriched representation power, then the multi-scale deconvolution feature fusion block is designed. We add the high-level features with rich semantic information to the low-level features via deconvolution fusion block. In order to locate aircraft of various scales more accurately, a series of aspect ratios of default boxes are set to better match aircraft shapes and combine predictions deduced from feature maps of different layers. The quantitative comparison analysis are carried out on the challenging UCAS-AOD data set. The experimental results demonstrate that the proposed method is accurate and robust for multi-scale aircraft detection, and achieves 94.8% AP(average precision) at the speed of 0.050 0 s/img with the input size 512×512 using a single Nvidia Titan Xp GPU.

Key words: remote sensing images, aircraft detection, feature fusion, multi-scale features

中图分类号:

P237

姚群力, 胡显, 雷宏. 基于多尺度融合特征卷积神经网络的遥感图像飞机目标检测[J]. 测绘学报, 2019, 48(10): 1266-1274.

YAO Qunli, HU Xian, Lei Hong. Aircraft detection in remote sensing imagery with multi-scale feature fusion convolutional neural networks[J]. Acta Geodaetica et Cartographica Sinica, 2019, 48(10): 1266-1274.

参考文献 25

[1]	CHENG Gong, ZHOU Peicheng, HAN Junwei. Learning rotation-invariant convolutional neural networks for object detection in VHR optical remote sensing images[J]. IEEE Transactions on Geoscience and Remote Sensing, 2016, 54(12):7405-7415.
[2]	HAN Junwei, ZHANG Dingwen, CHENG Gong, et al. Object detection in optical remote sensing images based on weakly supervised learning and high-level feature learning[J]. IEEE Transactions on Geoscience and Remote Sensing, 2015, 53(6):3325-3337.
[3]	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(12):119-132.
[4]	HE Chu, TU Mingxia, XIONG Dehui, et al. Adaptive component selection-based discriminative model for object detection in high-resolution SAR imagery[J]. ISPRS International Journal of Geo-Information, 2018, 7(2):72.
[5]	DIAO Wenhui, SUN Xian, ZHENG Xinwei, et al. Efficient saliency-based object detection in remote sensing images using deep belief networks[J]. IEEE Geoscience and Remote Sensing Letters, 2016, 13(2):137-141.
[6]	DU Bo, ZHANG Yuxiang, ZHANG Liangpei, et al. Beyond the sparsity-based target detector:a hybrid sparsity and statistics-based detector for hyperspectral images[J]. IEEE Transactions on Image Processing, 2016, 25(11):5345-5357.
[7]	LI Xuelong, MOU Lichao, LU Xiaoqiang. Scene parsing from an MAP perspective[J]. IEEE Transactions on Cybernetics, 2015, 45(9):1876-1886.
[8]	ZHANG Libao, ZHANG Yingying. Airport detection and aircraft recognition based on two-layer saliency model in high spatial resolution remote-sensing images[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2017, 10(4):1511-1524.
[9]	REN Shaoqing, HE Kaiming, GIRSHICK R, et al. Faster R-CNN:towards real-time object detection with region proposal networks[C]//Proceedings of the 28th International Conference on Neural Information Processing Systems. Montreal, Canada:ACM, 2015:91-99.
[10]	DAI Jifeng, LI Yi, HE Kaiming, et al. R-FCN:object detection via region-based fully convolutional networks[C]//Proceedings of the 30th International Conference on Neural Information Processing Systems. Barcelona, Spain:ACM, 2016:379-387.
[11]	REDMON J, FARHADI A. YOLO9000:better, faster, stronger[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Honolulu:IEEE, 2017:6517-6525.
[12]	LIU Wei, ANGUELOV D, ERHAN D, et al. SSD:single shot MultiBox detector[C]//Proceedings of the 14th European Conference On Computer Vision. Amsterdam, The Netherlands:Springer, 2016:21-37.
[13]	LIN T Y, DOLLÁR P, GIRSHICK R, et al. Feature pyramid networks for object detection[C]//Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition. Honolulu:IEEE, 2017:936-944.
[14]	HAN Xiaobing, ZHONG Yanfei, ZHANG Liangpei. An efficient and robust integrated geospatial object detection framework for high spatial resolution remote sensing imagery[J]. Remote Sensing, 2017, 9(7):666.
[15]	CAI Bowen, JIANG Zhiguo, ZHANG Haopeng, et al. Airport detection using end-to-end convolutional neural network with hard example mining[J]. Remote Sensing, 2017, 9(11):1198.
[16]	GUO Wei, YANG Wen, ZHANG Haijian, et al. Geospatial object detection in high resolution satellite images based on multi-scale convolutional neural network[J]. Remote Sensing, 2018, 10(1):131.
[17]	邓志鹏, 孙浩, 雷琳, 等. 基于多尺度形变特征卷积网络的高分辨率遥感影像目标检测[J]. 测绘学报, 2018, 47(9):1216-1227. DOI:10.11947/j.AGCS.2018.20170595. DENG Zhipeng, SUN Hao, LEI Lin, et al. Object detection in remote sensing imagery with multi-scale deformable convolutional networks[J]. Acta Geodaetica et Cartographica Sinica, 2018, 47(9):1216-1227. DOI:10.11947/j.AGCS.2018.20170595.
[18]	CHEN Zhong, ZHANG Ting, OUYANG Chao. End-to-end airplane detection using transfer learning in remote sensing images[J]. Remote Sensing, 2018, 10(1):139.
[19]	SHEN Zhiqiang, LIU Zhuang, LI Jianguo, et al. DSOD:learning deeply supervised object detectors from scratch[C]//Proceedings of 2017 IEEE International Conference on Computer Vision. Venice, Italy:IEEE, 2017:1937-1945.
[20]	BELL S, ZITNICK C L, BALA K, et al. Inside-outside net:detecting objects in context with skip pooling and recurrent neural networks[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas:IEEE, 2016:2874-2883.
[21]	KONG Tao, SUN Fuchun, YAO Anbang, et al. RON:reverse connection with objectness prior networks for object detection[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Honolulu:IEEE, 2017:5244-5252.
[22]	ZHU Haigang, CHEN Xiaogang, DAI Weiqun, et al. Orientation robust object detection in aerial images using deep convolutional neural network[C]//Proceedings of 2015 IEEE International Conference on Image Processing. Quebec City:IEEE, 2015:3735-3739.
[23]	LONG J, SHELHAMER E, DARRELL T. Fully convolutional networks for semantic segmentation[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Boston:IEEE, 2015:3431-3440.
[24]	梁华, 宋玉龙, 钱锋等. 基于深度学习的航空对地小目标检测[J]. 液晶与显示, 2018, 33(9):793-800. LIANG Hua, SONG Yulong, QIAN Feng, et al. Detection of small target in aerial photography based on deep learning[J]. Chinese Journal of Liquid Crystals and Displays, 2018, 33(9):793-800.
[25]	XIA Guisong, BAI Xiang, DING Jian, et al. DOTA:a large-scale dataset for object detection in aerial images[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City:IEEE, 2018:3974-3983.

基于多尺度融合特征卷积神经网络的遥感图像飞机目标检测

Aircraft detection in remote sensing imagery with multi-scale feature fusion convolutional neural networks

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献 25

相关文章 15

编辑推荐

Metrics

本文评价

[1]	杨必胜, 陈驰, 董震. 面向智能化测绘的城市地物三维提取[J]. 测绘学报, 2022, 51(7): 1476-1484.
[2]	梁哲恒, 黎宵, 邓鹏, 盛森, 姜福泉. 融合多尺度特征注意力的遥感影像变化检测方法[J]. 测绘学报, 2022, 51(5): 668-676.
[3]	叶发茂, 孟祥龙, 董萌, 聂运菊, 葛芸, 陈晓勇. 遥感图像蚁群算法和加权图像到类距离检索法[J]. 测绘学报, 2021, 50(5): 612-620.
[4]	周建伟, 吴一全. MRELBP特征、Franklin矩和SVM相结合的遥感图像建筑物识别方法[J]. 测绘学报, 2020, 49(3): 355-364.
[5]	李雪, 张力, 王庆栋, 艾海滨. 多时相遥感影像语义分割色彩一致性对抗网络[J]. 测绘学报, 2020, 49(11): 1473-1484.
[6]	余东行, 郭海涛, 张保明, 赵传, 卢俊. 级联卷积神经网络的遥感影像飞机目标检测[J]. 测绘学报, 2019, 48(8): 1046-1058.
[7]	赵生银, 安如, 朱美如. 联合像元-深度-对象特征的遥感图像城市变化检测[J]. 测绘学报, 2019, 48(11): 1452-1463.
[8]	李牧闲, 桂志鹏, 成晓强, 吴华意, 秦昆. 多核学习与用户反馈结合的WMS图层检索方法[J]. 测绘学报, 2019, 48(10): 1320-1330.
[9]	林文杰, 李玉, 赵泉华. 结合MST划分和RHMRF-FCM算法的高分辨率遥感图像分割[J]. 测绘学报, 2019, 48(1): 64-74.
[10]	连仁包, 王卫星, 李娟. 自适应圆形模板及显著图的高分辨遥感图像道路提取[J]. 测绘学报, 2018, 47(7): 950-958.
[11]	戴玉超, 张静, Fatih PORIKLI, 何明一. 深度残差网络的多光谱遥感图像显著目标检测[J]. 测绘学报, 2018, 47(6): 873-881.
[12]	胡根生, 周文利, 梁栋, 鲍文霞. 融合引导滤波和迁移学习的薄云图像中地物信息恢复算法[J]. 测绘学报, 2018, 47(3): 348-358.
[13]	杨俊涛, 康志忠. 多尺度特征和马尔可夫随机场模型的电力线场景点云分类法[J]. 测绘学报, 2018, 47(2): 188-197.
[14]	王峰萍, 王卫星, 薛柏玉, 曹霆, 高婷. GVF Snake与显著特征相结合的高分辨率遥感图像道路提取[J]. 测绘学报, 2017, 46(12): 1978-1985.
[15]	许夙晖, 慕晓冬, 赵鹏, 马骥. 利用多尺度特征与深度网络对遥感影像进行场景分类[J]. 测绘学报, 2016, 45(7): 834-840.