基于双注意力机制的遥感图像目标检测

doi:10.3969/j.issn.1006-2475.2020.08.001

摘要/Abstract

摘要： 针对遥感图像在复杂背景下小目标检测精度较低的问题，提出一种基于双注意力机制模型的SSD检测算法。该算法在前端特征提取网络中引入双注意力机制模型，强化低层特征图中小目标的有效特征信息并抑制冗余的语义信息，实现自适应特征学习；并在空间注意力模型中引入空洞卷积，保证卷积核感受野的同时减少了网络参数。引入Focal loss损失函数作为改进算法的分类损失函数，改善网络在训练过程中样本失衡的问题，增加正样本与难样本在训练时的权重比例，提升算法的检测性能。对遥感图像数据集NWPU VHR-10进行检测的结果表明，本文的改进算法在保证检测速度的同时提高了检测精度。与传统SSD算法相比，改进SSD算法的mAP提高了2.25个百分点，达到79.65%。

关键词: 深度学习, 目标检测, 特征提取, 双注意力机制模型, 空洞卷积, Focal loss损失函数

Abstract: Aiming at the problem of low accuracy of small target detection in remote sensing image under complex background, an new SSD detection algorithm based on dual attention mechanism model is proposed. The algorithm introduces the dual attention mechanism model in the front-end feature extraction network. It strengthens the effective feature information of small targets in the low-level feature map and suppresses the redundant semantic information to achieve adaptive feature learning. In addition, dilated convolution is introduced into the spatial attention model to ensure the sensitivity of convolution kernel and reduce the parameters of the network. The Focal loss function is introduced as the classification loss function of the improved algorithm to improve the imbalance of samples and increase the weight ratio of positive and difficult samples during training, it promotes the detection performance of the algorithms. The detection results of the remote sensing image data set NWPU VHR-10 show that the improved algorithm not only ensures the detection speed, but also improves the detection accuracy. Compared with the traditional SSD algorithm, the mAP of the improved SSD algorithm is increased by 2.25 percentage points to 79.65%.

Key words: deep learning, target detection, feature extraction, double attention mechanism model, dilated convolution, Focal loss function

中图分类号:

TP391.41

周幸, 陈立福. 基于双注意力机制的遥感图像目标检测[J]. 计算机与现代化, 2020, 0(08): 1-7.

ZHOU Xing, CHEN Li-fu. Object Detection of Remote Sensing Image Based on Dual Attention Mechanism[J]. Computer and Modernization, 2020, 0(08): 1-7.

参考文献［30］

［1］	VAN ETTEN A. You only look twice: Rapid multi-scale object detection in satellite imagery［DB/OL］. (2018-03-24)［2019-12-01］. https://arxiv.org/pdf/1805.09512.pdf.
［2］	郑胤,陈权崎,章毓晋. 深度学习及其在目标和行为识别中的新进展［J］. 中国图象图形学报, 2014,19(2):175-184.
［3］	曹林林,李海涛,韩颜顺,等. 卷积神经网络在高分遥感影像分类中的应用［J］. 测绘科学, 2016,41(9):170-175.
［4］	王金传,谭喜成,王召海,等. 基于Faster R-CNN深度网络的遥感影像目标识别方法研究［J］. 地球信息科学学报, 2018,20(10):1500-1508.
［5］	YAO X K, WAN L H, HOU H, et al. Airplane object detection in high resolution remote sensing imagery based on multi-structure convolutional neural network［J］. Computer Engineering, 2017,43(1):259-267.
［6］	徐逸之,姚晓婧,李祥,等. 基于全卷积网络的高分辨遥感影像目标检测［J］. 测绘通报, 2018(1):77-82.
［7］	ZOU Z X, SHI Z W. Ship detection in spaceborne optical image with SVD networks［J］. IEEE Transactions on Geoscience and Remote Sensing, 2016,54(10):5832-5845.
［8］	LI Q P, MOU L C, JIANG K Y, et al. Hierarchical region based convolution neural network for multiscale object detection in remote sensing images［C］// 2018 IEEE International Geoscience and Remote Sensing Symposium. 2018:4355-4358.
［9］	XIE H N, WANG T, QIAO M N, et al. Robust object detection for tiny and dense targets in VHR aerial images［C］// 2017 Chinese Automation Congress. 2017:6397-6401.
［10］	ZHANG W, WANG S H, THACHAN S, et al. Deconv R-CNN for small object detection on remote sensing images［C］// 2018 IEEE International Geoscience and Remote Sensing Symposium. 2018:2483-2486.
［11］	FU Y M, WU F G, ZHAO J S. Context-aware and depthwise-based detection on orbit for remote sensing image［C］// 2018 24th International Conference on Pattern Recognition. 2018:1725-1730.
［12］	姚远,姜志国,张浩鹏. 基于层次化分类器的遥感图像飞机目标检测［J］. 航天返回与遥感, 2014,35(5):88-94.
［13］	张宇,何楚,石博,等. 逐层特征选择的多层部件模型用于遥感图像飞机目标检测［J］. 武汉大学学报（信息科学版）, 2014,39(12):1406-1411.
［14］	ROSTEN E, PORTER R, DRUMMOND T. Faster and better: A machine learning approach to corner detection［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2010,32(1):105-119.
［15］	杨华,苏航,郑世宝. 大规模群体密度估计算法［J］. 电视技术, 2010,34(5):113-116.
［16］	PANDEY M, LAZEBNIK S. Scene recognition and weakly supervised object localization with deformable part-based models［C］// 2011 International Conference on Computer Vision. 2011:1307-1314.
［17］	AZIZPOUR H, LAPTEV I. Object detection using strongly-supervised deformable part models［C］// Proceedings of European Conference on Computer Vision. 2012:836-849.
［18］	GIRSHICK R, DONAHUE J, DARRELL T, et al. Rich feature hierarchies for accurate object detection and semantic segmentation［C］// Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2014:580-587.
［19］	GIRSHICKR. Fast R-CNN［C］// Proceedings of the IEEE International Conference on Computer Vision. 2015:1440-1448.
［20］	REN S Q, HE K M, 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. 2015:91-99.
［21］	REDMON J, DIVVALA S, GIRSHICK R, et al. You only look once: Unified, real-time object detection［C］// Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2016:779-788.
［22］	LIU W, ANGUELOV D, ERHAN D, et al. SSD: Single shot multibox detector［C］// European Conference on Computer Vision. 2016:21-37.
［23］	CHATFIELD K, SIMONYAN K, VEDALDI A, et al. Return of the Devil in the Details: Delving Deep into Convolutional Nets［DB/OL］.(2014-11-05)［2019-12-01］. https://arxiv.org/abs/1405.3531.
［24］	HE K M, ZHANG X Y, REN S Q, et al. Spatial pyramid pooling in deep convolutional networks for visual recognition［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2015,37(9):1904-1916.
［25］	REN S Q, HE K M, 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.
［26］	WOO S, PARK J, LEE J Y, et al. CBAM: Convolutional block attention module［C］// Proceedings of the European Conference on Computer Vision. 2018:3-19.
［27］	LIN M，CHEN Q，YAN S C. Network in network［DB/OL］. (2014-03-04)［2019-12-01］. https://arxiv.org/abs/1312.4400.
［28］	YU F, KOLTUN V. Multi-scale Context Aggregation by Dilated Convolutions［DB/OL］. (2016-04-30)［2019-12-01］. https://arxiv.org/abs/1511.07122.
［29］	LIN T Y, GOYAL P, GIRSHICK R, et al. Focal loss for dense object detection［C］// 2017 IEEE International Conference on Computer Vision. 2017:2999-3007.
［30］	IOFFE S, SZEGEDY C. Batch normalization: Accelerating deep network training by reducing internal covariate shift［DB/OL］. (2015-05-02)［2019-12-01］. https://arxiv.org/abs/1502.03167.

[1]	王秋忆, 周浩, 郑婷婷. 改进RetinaNet的电力设备目标检测方法[J]. 计算机与现代化, 2024, 0(01): 47-52.
[2]	胡崇佳, 刘金洲, 方立. 基于无监督域适应的室外点云语义分割[J]. 计算机与现代化, 2024, 0(01): 74-79.
[3]	林威. 基于自监督学习和数据回放的新闻推荐模型增量学习方法[J]. 计算机与现代化, 2023, 0(12): 1-6.
[4]	梁天恺, 黄康华, 刘凯航, 兰岚, 曾碧. 基于双向同态加密的深度联邦图片分类方法[J]. 计算机与现代化, 2023, 0(12): 36-40.
[5]	邱凯星, 冯广. 基于双重特征注意力的多标签图像分类模型[J]. 计算机与现代化, 2023, 0(12): 41-47.
[6]	张伯泉, 麦海鹏, 陈嘉敏, 逄锦聚. 基于高灰度值注意力机制的脑白质高信号分割[J]. 计算机与现代化, 2023, 0(12): 67-75.
[7]	马泽宇, 叶宁, 徐康, 王甦, 王汝传, . 基于FMCW雷达和ResNeSt-GRU的行为识别方法[J]. 计算机与现代化, 2023, 0(11): 101-107.
[8]	欧嘉城, 曾安, 金亮. 基于CP-YOLOX的冷冻电镜图像蛋白质目标检测算法[J]. 计算机与现代化, 2023, 0(11): 113-119.
[9]	闫子贤, 董宝良, 唐思谜. 针对复杂背景下低分辨率舰船目标的改进YOLOv7算法[J]. 计算机与现代化, 2023, 0(11): 120-126.
[10]	李延满, 王必恒, 赵羚焱. 基于轻量化YOLOv5的安全帽检测[J]. 计算机与现代化, 2023, 0(10): 59-64.
[11]	黎世达, 项剑文. 一种提高图像识别模型鲁棒性的弱化强化方法[J]. 计算机与现代化, 2023, 0(10): 70-76.
[12]	沈加炜, 陆一鸣, 陈晓艺, 钱美玲, 陆卫忠, . 基于深度学习的人体行为检测方法研究综述[J]. 计算机与现代化, 2023, 0(09): 1-9.
[13]	顾成伟, 丁勇, 李登华. 基于计算机视觉的工业厂区人员安全警戒系统[J]. 计算机与现代化, 2023, 0(09): 20-26.
[14]	杨柳青, 王冲. 基于极大熵的Web服务资源个性化推荐方法[J]. 计算机与现代化, 2023, 0(09): 32-37.
[15]	刘禅奕, 黄丹, 薛林雁, 王涛, 朱桃, . 改进EfficientNet网络的COVID-19 X光分类[J]. 计算机与现代化, 2023, 0(09): 94-99.