李朝奎,方军,吴馁,宋璟毓,周倩,周青蓝

• 1:湖南科技大学地理空间信息技术国家地方联合工程实验室
• 2:湖南科技大学地理空间信息湖南省工程实验室
• 3:中国船舶工业系统工程研究院

摘要(Abstract)：

针对高分辨率遥感影像道路提取过程中,深度学习方法较传统提取方法虽可有效地提高地物提取的精度,但需要大量样本训练,消耗较多计算资源,且大量高质量训练样本难以获得等问题,该文提出了一种将支持向量机(SVM)与卷积神经网络(CNN)结合的适用于小样本的道路提取混合模型,该模型能够在小样本训练时保证道路提取的精度。采用数据增强、正则化等方法优化训练策略,丰富小样本道路特征库,设计结合SVM的深度卷积神经网络结构来提取道路在影像中的高维特征,降低模型计算量,减少计算时间。以谷歌高分辨率遥感影像作为实验数据,用不同训练样本量来训练模型并验证道路提取的精度;同时,将该文提出的方法与逻辑回归(LR)模型、光谱结合SVM模型以及VGG16深度学习模型进行了道路提取效果的对比分析。结果表明:该文倡导模型方法在小样本情况下可以提取较高精度道路;与其他3种方法比较,该文方法能够快速构建与训练模型,在满足精度要求的同时,极大地提高了高分辨率遥感影像道路提取的效率,为道路数据的快速更新与变化检测提供了新的技术支持。

关键词(KeyWords)： 深度卷积神经网络;SVM;小样本;高分辨率遥感影像;道路提取

Abstract:

Keywords:

基金项目(Foundation): 国家重点研发计划项目课题(2017YFB0503802);; 国家自然科学基金项目(41571374)

作者(Author): 李朝奎,方军,吴馁,宋璟毓,周倩,周青蓝

DOI: 10.16251/j.cnki.1009-2307.2020.04.013

参考文献(References)：

• [1] 史文中,朱长青,王昱.从遥感影像提取道路特征的方法综述与展望[J].测绘学报,2001,30(3):257-262.(SHI Wenzhong,ZHU Changqing,WANG Yu.Road feature extraction from remotely sensed image:review and prospects[J].Acta Geodaetica et Cartographica Sinica,2001,30(3):257-262.)
• [2] 曹云刚,王志盼,杨磊.高分辨率遥感影像道路提取方法研究进展[J].遥感技术与应用,2017,32(1):20-26.(CAO Yungang,WANG Zhipan,YANG Lei.Advances in method on road extraction from high resolution remote sensing images[J].Remote Sensing Technology and Application,2017,32(1):20-26.)
• [3] 吴亮,胡云安.遥感图像自动道路提取方法综述[J].自动化学报,2010,36(7):912-922.(WU Liang,HU Yun’an.A survey of automatic road extraction from remote sensing images[J].Acta Automatica Sinica,2010,36(7):912-922.)
• [4] 林祥国,张继贤,李海涛,等.基于T型模板匹配半自动提取高分辨率遥感影像带状道路[J].武汉大学学报(信息科学版),2009,34(3):293-296.(LIN Xiangguo,ZHANG Jixian,LI Haitao,et al.Semi-automatic extraction of ribbon road from high resolution remotely sensed imagery by a T-shaped template matching[J].Geomatics and Information Science of Wuhan University,2009,34(3):293-296.)
• [5] YI W B,CHEN Y H,TANG H,et al.Experimental research on urban road extraction from high-resolution RS images using probabilistic topic models[C]//Geoscience and Remote Sensing Symposium.Piscataway:IEEE,2010:445-448.
• [6] CHEN H,YIN L,MA L L.Research on road information extraction from high resolution imagery based on global precedence[C]//International Workshop on Earth Observation and Remote Sensing Applications.Piscataway:IEEE,2014:151-155.
• [7] 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.)
• [8] 何小飞,邹峥嵘,陶超,等.联合显著性和多层卷积神经网络的高分影像场景分类[J].测绘学报,2016,45(9):1073-1080.(HE Xiaofei,ZOU Zhengrong,TAO Chao,et al.Combined saliency with multi-convolutional neural network for high resolution remote sensing scene classification[J].Acta Geodaetica et Cartographica Sinica,2016,45(9):1073-1080.)
• [9] 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.(CAO Linlin,LI Haitao,HAN Yanshun,et al.Application of convolutional neural networks in classification of high resolution remote sensing imagery[J].Science of Surveying and Mapping,2016,41(9):170-175.)
• [10] 欧阳颖卉,林翬,李树涛.基于卷积神经网络的光学遥感图像船只检测[J].包装工程,2016,37(15):1-6.(OUYANG Yinghui,LIN Hui,LI Shutao.Convolutional neural network based on ship detection in optical remote sensing image[J].Packaging Engineering,2016,37(15):1-6.)
• [11] ZHONG Z L,LI J,CUI W H,et al.Fully convolutional networks for building and road extraction:preliminary results[C]//Geoscience and Remote Sensing Symposium.Piscataway:IEEE,2016:1591-1594.
• [12] BITTNER K,CUI S,REINARTZ P.Building extraction from remote sensing data using fully convolutional networks[J].International Archives of the Photogrammetry,Remote Sensing and Spatial Information Sciences,2017,XLII-1/W1:481-486.
• [13] 张伟,郑柯,唐娉,等.深度卷积神经网络特征提取用于地表覆盖分类初探[J].中国图象图形学报,2017,22(8):1144-1153.(ZHANG Wei,ZHENG Ke,TANG Ping,et al.Land cover classification with features extracted by deep convolutional neural network[J].Journal of Image and Graphics,2017,22(8):1144-1153.)
• [14] 叶雪娜.基于卷积神经网络的遥感图像道路提取研究[D].西安:陕西师范大学,2017.(YE Xuena.Research on road extraction of remote sensing image based on convolutional neural network[D].Xi’an:Shaanxi Normal University,2017.)
• [15] 刘志强,夏威,彭红春.基于深度学习的高分辨率遥感影像道路提取[C]//高分辨率对地观测学术年会.[出版地不详]:[出版者不详],2017.(LIU Zhiqiang,XIA Wei,PENG Hongchun.Road extraction based on deep learning and high resolution remote sensing image[C]//High Resolution Earth Observation Conference.[S.l.]:[s.n.],2017.)
• [16] 张永宏,夏广浩,阚希,等.基于全卷积神经网络的多源高分辨率遥感道路提取[J].计算机应用,2018,38(7):2070-2075.(ZHANG Yonghong,XIA Guanghao,KAN Xi,et al.Road extraction from multi-source high resolution remote sensing image based on fully convolutional neural network[J].Journal of Computer Applications,2018,38(7):2070-2075.)
• [17] WEI Y,WANG Z,XU M.Road structure refined CNN for road extraction in aerial image[J].IEEE Geoscience & Remote Sensing Letters,2017,14(5):709-713.
• [18] XU B,WANG N Y,CHEN T Q,et al.Empirical evaluation of rectified activations in convolutional network[J/OL].[2017-12-28].https://arxiv.org/pdf/1505.00853v1.pdf.
• [19] 周飞燕,金林鹏,董军.卷积神经网络研究综述[J].计算机学报,2017,40(6):1229-1251.(ZHOU Feiyan,JIN Linpeng,DONG Jun.Review of convolutional neural network[J].Chinese Journal of Computers,2017,40(6):1229-1251.)
• [20] 余永维,殷国富,殷鹰,等.基于深度学习网络的射线图像缺陷识别方法[J].仪器仪表学报,2014,35(9):2012-2019.(YU Yongwei,YIN Guofu,YIN Ying,et al.Defect recognition for radiographic image based on deep learning network[J].Chinese Journal of Scientific Instrument,2014,35(9):2012-2019.)