Please wait a minute...
img

官方微信

遥感技术与应用  2020, Vol. 35 Issue (6): 1377-1385    DOI: 10.11873/j.issn.1004-0323.2020.6.1377
数据与图像处理     
使用贝叶斯优化对遥感影像目标进行精确定位
柴栋1(),许夙晖1,2(),罗畅3,鲁彦辰4
1.空军研究院,北京 100085
2.中国人民解放军78102部队,四川 成都 610031
3.中国人民解放军78092部队,四川 成都 610031
4.中国人民解放军96946部队,北京 102202
Object Accurate Localization of Remote Sensing Image based on Bayesian Optimization
Dong Chai1(),Suhui Xu1,2(),Chang Luo3,Yanchen Lu4
1.Beijing Aviation Engineering Technology Research Center,Beijing 100076,China
2.The Chinese People's Liberation Army(78102),Chengdu,Sichuan 610031,China
3.The Chinese People's Liberation Army(78092),Chengdu,Sichuan 610031,China
4.The Chinese People's Liberation Army(96946),Beijing 102202,China
 全文: PDF(5080 KB)   HTML
摘要:

针对大尺寸遥感影像目标检测中检测边框不精确的问题,提出使用高斯过程贝叶斯优化对遥感影像中的目标进行精确检测与定位。研究分为两个阶段,第一阶段使用基于边缘信息的EdgeBoxes算法对大尺寸遥感影像进行目标候选区域的选取,用分类器得到初始检测结果;为了得到更加准确的边框,在第二阶段,基于高斯过程的贝叶斯优化对每个目标的边框进行微调:①以目标初始边框为基准,在其周围选取与初始边框相交的边框集合,并得到一个高斯过程分布;②使用贝叶斯优化估计出下一个边框,并将其加入边框集;③求分类器对所有边框的得分,得分最高的边框作为下次迭代的基准边框;④重复若干次贝叶斯优化后得到最终的边框。实验结果表明:EdgeBoxes方法以较少的候选框可以得到较大的召回率,使用高斯过程的贝叶斯优化可以明显地提高检测边框的精度。

关键词: 遥感影像目标检测目标精确定位区域候选框高斯过程贝叶斯优化    
Abstract:

To solve the issues of inaccurate bounding box in large-scale remote sensing image object detection, an accurate object detection and localization appoarch of remote sensing image based on Bayesian Optimization is proposed. The method consists of two stages: In the first stage, the EdgeBoxes which is based on edges information is adopted to generate object proposals. The classifier is applied to get initial object detection result. To obtain more accurate bounding box, a bayesian optimization based on gaussian process is applied to fine-tune the bounding box around each object in the second stage. Firstly, a set of boxes that intersect the initial bounding box around each initial box is selected to form a gaussian process. Secondly, a new bounding box is estimated through bayesian optimization and added to the set of boxes. Thirdly, the score of each box is calculated by the classifier, and the box with the highest score is set as the base box in the next iteration. At last, the bayesian optimization process is repeatedand and final bounding boxes is obtained. Experiments demonstrate the EdgeBoxes method can achive a better recall evaluation with less number of propsals. The bayesian optimization based on gaussian process can significantly improve the localization accuracy of the detection bounding box.

Key words: Remote sensing image    Object detection    Object accurate locolization    Region proposal    Gaussian process    Bayesian optimization
收稿日期: 2019-12-09 出版日期: 2021-01-26
ZTFLH:  TP751  
基金资助: 军内某科研项目
通讯作者: 许夙晖     E-mail: chaibaodong@126.com;xu_suhui@163.com
作者简介: 柴栋(1988-),男,河南博爱人,博士,助理研究员,主要从事飞行器总体、人工智能研究。E?mail: chaibaodong@126.com
服务  
把本文推荐给朋友
加入引用管理器
E-mail Alert
RSS
作者相关文章  
柴栋
许夙晖
罗畅
鲁彦辰

引用本文:

柴栋,许夙晖,罗畅,鲁彦辰. 使用贝叶斯优化对遥感影像目标进行精确定位[J]. 遥感技术与应用, 2020, 35(6): 1377-1385.

Dong Chai,Suhui Xu,Chang Luo,Yanchen Lu. Object Accurate Localization of Remote Sensing Image based on Bayesian Optimization. Remote Sensing Technology and Application, 2020, 35(6): 1377-1385.

链接本文:

http://www.rsta.ac.cn/CN/10.11873/j.issn.1004-0323.2020.6.1377        http://www.rsta.ac.cn/CN/Y2020/V35/I6/1377

图1  本研究解决的问题
图2  遥感影像目标初步检测框架
图3  基于高斯过程贝叶斯优化的目标边框微调
图4  数据集样例
数据集用途张数目标数尺寸
RSOD-airplane训练正样本4465 0611 044×915
SanFrancisco 机场测试图4774 800×2 589
Denver机场测试图41034 800×2 589
LosAngeles机场测试图2394 800×2 589
其他训练负样本304 800×2 589
表1  数据集信息
参数名称αβMinScoreMinBoxAreaMaxBoxes
设定值0.010.750.012 00020 000
表2  EdgeBoxes参数设置
图5  两幅测试图像进行优化前后的检测结果与真实目标框对比
图6  4个目标的得分和IOU与迭代次数的关系曲线
图7  6种方法对8幅测试图像检测结果评价
SS-ISSGP-ISS-IISSGP-IIEdgeEdgeGP
召回率(IOU≥0.5)150/219160/219144/219146/219210/219214/219
召回率(IOU≥0.7)78/219107/21980/21992/219125/219161/219
平均值 IOU0.620 20.653 30.621 20.652 20.700 30.743 2
表3  6种方法对测试图像检测结果评价
1 Wang Yingjie, Zhang Qiao, Zhang Yanmei, et al. Oil Tank Detection from Remote Sensing Images based on Deep Convolutional Neural Network[J]. Remote Sensing Technology and Application, 2019, 34(4): 727-735.
1 王颖洁, 张荞, 张艳梅, 等. 基于深度卷积神经网络的油罐目标检测研究[J]. 遥感技术与应用, 2019, 34(4): 727-735.
2 Cheng G, Han J. A Survey on Object Detection in Optical Remote Sensing Images[J]. ISPRS Journal of Photogrammetry & Remote Sensing, 2016, 117:11-28. doi: 10. 1109/TGRS. 2014. 2374218.
doi: 10. 1109/TGRS. 2014. 2374218
3 Yang Bisheng,Zong Zeliang,Chen Chi,et al. Real Time Approach for Underground Objects Detection from Vehicle-borne Ground Penetrating Radar[J]. Acta Geodaetica et Cartographica Sinica, 2020, 49(7):874-883.
3 杨必胜, 宗泽亮, 陈驰, 等. 车载探地雷达地下目标实时探测法[J]. 测绘学报, 2020, 49(7):874-882.
4 Yang Rui, Qi Yuan, Su Yang. U-Net Neural Networks and Its Application in High Resolution Satellite Image Classification. Remote Sensing Technology and Application[J], 2020, 35(4): 767-774.
4 杨瑞, 祁元, 苏阳. 深度学习U-Net方法及其在高分辨卫星影像分类中的应用. 遥感技术与应用[J], 2020, 35(4): 767-774.
5 Zheng Xin, Pan Bin, Zhang Jian. Power Tower Detection in Remote Sensing Imagery based on Deformable Network and Transfer Learning[J]. Acta Geodaetica et Cartographica Sinica, 2020, 49(8):1042-1050.
5 郑鑫, 潘斌, 张健. 可变形网络与迁移学习相结合的电力塔遥感影像目标检测法[J]. 测绘学报, 2020, 49(8):1042-1050.
6 Han J, Zhang D, Cheng G, 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. doi: 10. 1109/TGRS. 2014. 2374218.
doi: 10. 1109/TGRS. 2014. 2374218
7 Zhang F, Du B, Zhang L, et al. Weakly Supervised Learning based on Coupled Convolutional Neural Networks for Aircraft Detection[J]. IEEE Transactions on Geoscience and Remote Sensin, 2016, 54(9): 5553. doi: 10. 1109/TGRS. 2016. 2569141.
doi: 10. 1109/TGRS. 2016. 2569141
8 Long Y, Gong Y, Xiao Z, et al. Accurate Object Localization in Remote Sensing Images based on Convolutional Neural Networks[J]. IEEE Transactions on Geoscience and Remote Sensing, 2017, 55(5): 2486-2498. doi: 10. 1109/TGRS. 2016. 2645610.
doi: 10. 1109/TGRS. 2016. 2645610
9 Tang J, Deng C, Huang G B, et al. Compressed-Domain Ship Detection on Spaceborne Optical Image Using Deep Neural Network and Extreme Learning Machine[J]. IEEE Transactions on Geoscience and Remote Sensing, 2014, 53(3):1174-1185. doi: 10. 1109/TGRS. 2014. 2335751.
doi: 10. 1109/TGRS. 2014. 2335751
10 Yang Y, Zhuang Y, Bi F, et al. M-FCN: Effective Fully Convolutional Network-based Airplane Detection Framework[J]. IEEE Geoscience and Remote Sensing Letters, 2017, 14(8): 1293-1297. doi: 10. 1109/LGRS. 2017. 2708722.
doi: 10. 1109/LGRS. 2017. 2708722
11 Dalal N, Triggs B. Histograms of Oriented Gradients for Human Detection[C]∥ Computer Vision and Pattern Recognition, 2005. CVPR2005.
11 IEEE Computer Society Conference on. IEEE, 2005, 1: 886-893.
12 Hosang J, Benenson R, Dollár P, et al. What Makes for Effective Detection Proposals?[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2016, 38(4): 814-830. doi: 10. 1109/TPAMI. 2015. 2465908.
doi: 10. 1109/TPAMI. 2015. 2465908
13 Uijlings J R R , Sande K E A V D , Gevers T , et al. Selective Search for Object Recognition[J]. International Journal of Computer Vision, 2013, 104(2):154-171. doi:10. 1007/s11263-013-0620-5.
doi: 10. 1007/s11263-013-0620-5
14 Cheng M M, Zhang Z, Lin W Y, et al. BING: Binarized Normed Gradients for Objectness Estimation at 300fps[C]∥ IEEE Conference on Computer Vision and Pattern Recognition. IEEE Computer Society, 2014:3286-3293. doi: 10. 1109/CVPR. 2014. 414.
doi: 10. 1109/CVPR. 2014. 414
15 Erhan D, Szegedy C, Toshev A, et al. Scalable Object Detection Using Deep Neural Networks[C]∥ Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2014: 2147-2154. doi: 10. 1109/CVPR. 2014. 276.
doi: 10. 1109/CVPR. 2014. 276
16 Zitnick C L, Dollár P. Edge Boxes: Locating Object Proposals from edges[C]∥ European Conference on Computer Vision. Springer, Cham, 2014: 391-405. doi: 10. 1007/978-3-319-10602-1_26.
doi: 10. 1007/978-3-319-10602-1_26
17 Simonyan K, Zisserman A. Very Deep Convolutional Networks for Large-scale Image Recognition[J]. Computerence, 2015. doi:arxiv.org/abs/1409.1556.
doi: arxiv.org/abs/1409.1556
18 Lizotte D J. Practical Bayesian Optimization[M]. Canadian: University of Alberta, 2008.
19 Lu Pengjie, Xu Dalu, Ren Fu, et al. Auto‐detection and Hiding of Sensitive Targets in Emergency Mapping based on Remote Sensing Data[J]. Geomatics and Information Science of Wuhan University, 2020, 45(8):1263-1272.
19 鲁鹏杰, 许大璐, 任福, 等. 应急遥感制图中敏感目标自动检测与隐藏方法[J]. 武汉大学学报(信息科学版), 2020, 45(8):1263-1272.
[1] 杨瑞,祁元,苏阳. 深度学习U-Net方法及其在高分辨卫星影像分类中的应用[J]. 遥感技术与应用, 2020, 35(4): 767-774.
[2] 王卓,闫浩文,禄小敏,冯天文,李亚珍. 一种改进U-Net的高分辨率遥感影像道路提取方法[J]. 遥感技术与应用, 2020, 35(4): 741-748.
[3] 连喜红,祁元,王宏伟,张金龙,杨瑞. 基于面向对象的青海湖环湖区居民地信息自动化提取[J]. 遥感技术与应用, 2020, 35(4): 775-785.
[4] 曹小杰,江洪,张兆明,何国金,赵晶晶. Landsat8卫星影像大气校正及其植被指数与SEVI性能比较[J]. 遥感技术与应用, 2019, 34(5): 983-991.
[5] 王颖洁,张荞,张艳梅,蒙印,郭文. 基于深度卷积神经网络的油罐目标检测研究[J]. 遥感技术与应用, 2019, 34(4): 727-735.
[6] 丛铭, 段晨曦, 许妙忠, 陶翊婷. 基于格式塔形状分析的高分辨率遥感影像道路提取[J]. 遥感技术与应用, 2019, 34(3): 595-601.
[7] 李淑敏, 冯权泷, 梁其椿, 张学庆. 基于深度学习的国产高分遥感影像飞机目标自动检测[J]. 遥感技术与应用, 2018, 33(6): 1095-1102.
[8] 冀新莹, 韦玉春, 王问尧, 方宏. 城镇区域高分辨率遥感影像地表覆盖变化检测的误差分析[J]. 遥感技术与应用, 2018, 33(5): 932-941.
[9] 周子勇. 运用经验模态分解和压缩感知方法进行遥感影像超分辨率重建[J]. 遥感技术与应用, 2018, 33(1): 96-102.
[10] 周为峰,曹利,李小恕,程田飞. 沿海牡蛎养殖的WorldView-2影像融合方法评价[J]. 遥感技术与应用, 2018, 33(1): 103-109.
[11] 熊伟,徐永力,姚力波,崔亚奇. 基于SVM的高分辨率SAR图像舰船目标检测算法[J]. 遥感技术与应用, 2018, 33(1): 119-127.
[12] 林齐根,邹振华,祝瑛琦,王瑛. 基于光谱、空间和形态特征的面向对象滑坡识别[J]. 遥感技术与应用, 2017, 32(5): 931-937.
[13] 吕利利,颉耀文,董龙龙. 基于不同地形校正模型的影像反射率对比分析[J]. 遥感技术与应用, 2017, 32(4): 751-759.
[14] 吕利利,颉耀文,黄晓君,张秀霞,李汝嫣. 基于CART决策树分类的沙漠化信息提取方法研究[J]. 遥感技术与应用, 2017, 32(3): 499-506.
[15] 王艳楠,王健健,龚健新,袁帅,刘辉,罗文. 基于环境卫星数据的沿海滩涂地物类型分类的随机森林方法[J]. 遥感技术与应用, 2016, 31(6): 1107-1113.