U型卷积神经网络的ZY-3影像道路提取方法Road extraction from ZY-3 remote sensing image based on U-Net like convolution architecture
郭正胜;李参海;王智敏;
摘要(Abstract)：
针对经典全卷积神经网络在池化和上采样过程中造成图像分辨率不断下降以及对各个像素进行分类时忽略了像素之间的关系,导致提取道路比较模糊和平滑的问题。该文提出一种基于U型卷积网络的ZY-3道路提取方法。首先,参考医学图像分割领域表现突出的U-Net模型,采用对称式网络结构将低级细节信息与高级语义信息相结合,提高道路的初提取精度;其次考虑到卷积神经网络对百万量级的参数优化程度相对不足,采用集成学习的方法,通过变更权重获得若干个模型进行融合,进一步提升了道路提取的精度;最后,通过使用形态学开运算完成孔洞的去除等工作。实验结果表明,该文方法的提取结果在不同实验区域中平均准确度达到了95%以上,显著优于基于经典全卷积网络模型、基于纹理与形状特征提取道路的方法。
关键词(KeyWords)： 道路提取;ZY-3影像;卷积神经网络;U型卷积网络;集成学习
基金项目(Foundation): 国家重点研发计划项目(2016YFB0501403)
作者(Authors): 郭正胜;李参海;王智敏;
DOI: 10.16251/j.cnki.1009-2307.2020.04.009
参考文献(References)：
- [1] 张永宏,何静,阚希,等.遥感图像道路提取方法综述[J].计算机工程与应用,2018,54(13):1-10,51.(ZHANG Yonghong,HE Jing,KAN Xi,et al.Summary of road extraction methods for remote sensing images[J].Computer Engineering and Applications,2018,54(13):1-10,51.)
- [2] 左娟,李勇军.结合纹理与形状特征的高分辨率遥感影像道路提取[J].测绘,2013,36(3):111-113.(ZUO Juan,LI Yongjun.Road extraction from high resolution remote sensing image based on texture and shape feature[J].Surveying and Mapping,2013,36(3):111-113.)
- [3] 赵文智,雒立群,郭舟,等.光谱特征分析的城市道路提取[J].光谱学与光谱分析,2015,35 (10):2814-2819.(ZHAO Wenzhi,LUO Liqun,GUO Zhou,et al.Road extraction in remote sensing images based on spectral and edge analysis[J].Spectroscopy & Spectral Analysis,2015,35(10):2814-2819.)
- [4] GOMEZ R B.Hyperspectral imaging:a useful technology for transportation analysis[J].Optical Engineering,2002,41(41):2137-2143.
- [5] ZHOU B,LAPEDRIZA A,XIAO J,et al.Learning deep features for scene recognition using places database[C]//International Conference on Neural Information Processing Systems.Cambridge,MA,USA:MIT Press,2014:487-495.
- [6] REN S,HE K,GIRSHICK R,et al.Faster R-CNN:towards real-time object detection with region proposal networks[J].IEEE Transactions on Pattern Analysis & Machine Intelligence,2017,39(6):1137-1149.
- [7] ZHENG Y,LIU Q,CHEN E H,et al.Exploiting multi-channels deep convolutional neural networks for multivariate time series[J].Frontiers of Computer Science,2016,10(1):96-112.
- [8] LI P,ZANG Y,WANG C,et al.Road network extraction via deep learning and line integral convolution[C]//Geoscience and Remote Sensing Symposium.[S.l.]:IEEE,2016:1599-1602.
- [9] 刘笑,王光辉,杨化超,等.全卷积神经网络遥感影像道路提取方法[J].遥感信息,2018,33(1):69-75.(LIU Xiao,WANG Guanghui,YANG Huachao,et al.Road extraction from remote sensing image based on fully convolutional networks[J].Remote Sensing Information,2018,33(1):69-75.)
- [10] RONNEBERGER O,FISCHER P,BROX T.U-Net:convolutional networks for biomedical image segmentation[C]//International Conference on Medical Image Computing and Computer- Assisted Intervention.Cham:Springer,2015:234-241.
- [11] LONG J,SHELHAMER E,DARRELL T.Fully convolutional networks for semantic segmentation[J].IEEE Transactions on Pattern Analysis & Machine Intelligence,2014,39(4):640-651.
- [12] BREIMAN L.Bagging predictors[J].Machine Learning,1996,24(2):123-140.
- [13] 王昆,万幼川,屈颖.融合纹理和形状特征的高分辨率遥感影像道路提取[J].遥感信息,2010(5):7-11.(WANG Kun,WAN Youchuan,QU Ying.Road extraction from high-resolution remote sensing images based on texture and shape features[J].Remote Sensing Information,2010(5):7-11.)
- [14] JIA Y,SHELHAMER E,DONAHUE J,et al.Caffe:convolutional architecture for fast feature embedding[C]//Proceedings of the 22nd ACM international conference on Multimedia.New York:ACM,2014:675-678.
- [15] CHEN Q,SUN Q S,HENG P A,et al.A double-threshold image binarization method based on edge detector[J].Pattern Recognition,2008,41(4):1254-1267.