How to accurately extract large-scale urban land? Establishment of an improved fully convolutional neural network model

Abstract

Realizing accurate perception of urban boundary changes is conducive to the formulation of regional development planning and researches of urban sustainable development. In this paper, an improved fully convolution neural network was provided for perceiving large-scale urban change, by modifying network structure and updating network strategy to extract richer feature information, and to meet the requirement of urban construction land extraction under the background of large-scale low-resolution image. This paper takes the Yangtze River Economic Belt of China as an empirical object to verify the practicability of the network, the results show the extraction results of the improved fully convolutional neural network model reached a precision of kappa coefficient of 0.88, which is better than traditional fully convolutional neural networks, it performs well in the construction land extraction at the scale of small and medium-sized cities.

Keywords

improved fully convolutional neural network / remote sensing image classification / city boundary / precision evaluation.

Boling YIN is a M.S. candidate at the Chongqing Jiaotong University and his major is cartography and geographic information system. His research interest is the 3S technology integration and applicationEmail: [e-mail]

Dongjie GUAN is a professor at the Chongqing Jiaotong University. She received her PhD Degree in Environmental engineering from the University of Kitakyushu in 2009. She is the author of 55 papers. Her current research interests include land simulation, remote sensing analysis, and ecological compensation.E mail: [e-mail]

Yuxiang ZHANG is a M.S. candidate at the Chongqing Jiaotong University and his major is cartography and geographic information system. His research interest is the ecological compensation.Email: [e-mail]

He XIAO is a researcher at the Chongqing Geomatics and Remote Sensing Center. His research interest is the remote sensing and geographic information system.Email: [e-mail]

Lidan CHENG is a researcher at the Chongqing Geomatics and Remote Sensing Center. Her research interest is the ecological security assessment.Email: [e-mail]

Jiameng CAO is a M.S. candidate at the Chongqing Jiaotong University and her major is physical geography. Her research interest is the ecological risk assessment.Email: [e-mail]

Xiangyuan SU is a M.S. candidate at the Chongqing Jiaotong University and her major is human geography. Her research interest is the Eco-compensation.Email: [e-mail]

References

[1] BarthR, IJsselmuidenJ, HemmingJ, Van HentenE J. ( 2019). Synthetic bootstrapping of convolutional neural networks for semantic plant part segmentation. Comput Electron Agric, 161: 291– 304

[2] ChinaStatistical Yearbook ( 2019). China 2010 Population Census Data. Beijing: China Statistics Press

[3] DengZ, SunH, ZhouS, ZhaoJ, LeiL, ZouH. ( 2018). Multi-scale object detection in remote sensing imagery with convolutional neural networks. ISPRS J Photogramm Remote Sens, 145: 3– 22

[4] DingP, ZhangY, DengW J, JiaP, KuijperA. ( 2018). A light and faster regional convolutional neural network for object detection in optical remote sensing images. ISPRS J Photogramm Remote Sens, 141: 208– 218

[5] DuchiJ, HazanE, SingerY. ( 2011). Adaptive subgradient methods for online learning and stochastic optimization. J Mach Learn Res, 12( 7): 257– 269

[6] FloodN, WatsonF, CollettL. ( 1998). Using a U-net convolutional neural network to map woody vegetation extent from high resolution satellite imagery across Queensland, Australia. ITC J, 82: 101897

[7] FuG, LiuC, ZhouR, SunT, ZhangQ. ( 2017). Classification for high resolution remote sensing imagery using a fully convolutional network. Remote Sens-Basel, 9( 6): 1– 21

[8] FuY Y, LiuK K, ShenZ Q, DengJ S, GanM Y, LiuX G, LuD M, WangK. ( 2019). Mapping impervious surfaces in town–rural transition belts using China’s GF-2 imagery and object-based deep CNNs. Remote Sens (Basel), 11( 3): 280

[9] GebrehiwotA, Hashemi-BeniL, ThompsonG, KordjamshidiP, LanganT E. ( 2019). Deep convolutional neural network for flood extent mapping using unmanned aerial vehicles data. Sensors (Basel), 19( 7): 1486

[10] GongP, LiuH, ZhangM N, LiC, WangJ, HuangH, ClintonN, JiL, LiW, BaiY, ChenB, XuB, ZhuZ, YuanC, Ping SuenH, GuoJ, XuN, LiW, ZhaoY, YangJ, YuC, WangX, FuH, YuL, DronovaI, HuiF, ChengX, ShiX, XiaoF, LiuQ, SongL. ( 2019). Stable classification with limited sample: transferring a 30-m resolution sample set collected in 2015 to mapping 10-m resolution global land cover in 2017. Sci Bull (Beijing), 64( 6): 370– 373

[11] GuoH N, ShiQ, MarinoniA, DuB, ZhangL P. ( 2021). Deep building footprint update network: a semi-supervised method for updating existing building footprint from bi-temporal remote sensing images. Remote Sens Environ, 264: 112589

[12] HanZ M, DianY Y, XiaH, ZhouJ J, JianY F, YaoC H, WangX, LiY. ( 2020). Comparing fully deep convolutional neural networks for land cover classification with high-spatial-resolution Gaofen-2 images. ISPRS Int J Geoinf, 9( 8): 478

[13] HeC Y, LiuZ F, GouS Y, ZhangQ F, ZhangJ S, XuL L. ( 2019). Detecting global urban expansion over the last three decades using a fully convolutional network. Environ Res Lett, 14( 3): 034008

[14] HeD, ShiQ, LiuX P, ZhongY F, ZhangX C. ( 2021). Deep subpixel mapping based on semantic information modulated network for urban land use mapping. IEEE T Geosci Remote, pp( 99): 1– 19

[15] HuY, ZhangQ, ZhangY, YanH. ( 2018). A deep convolution neural network method for land cover mapping: a case study of Qinhuangdao, China. Remote Sens (Basel), 10( 12): 2053– 2069

[16] HuangB, ZhaoB, SongY. ( 2018). Urban land-use mapping using a deep convolutional neural network with high spatial resolution multispectral remote sensing imagery. Remote Sens Environ, 214: 73– 86

[17] JeanN, BurkeM, XieM, DavisW M, LobellD B, ErmonS. ( 2016). Combining satellite imagery and machine learning to predict poverty. Science, 353( 6301): 790– 794

[18] JiS P, ZhangC, XuA J, ShiY, DuanY L. ( 2018). 3D convolutional neural networks for crop classification with multi-temporal remote sensing images. Remote Sens (Basel), 10( 1): 75

[19] JiS, WeiS, LuM. ( 2019). Fully convolutional networks for multisourcebuilding extraction from an open aerial and satellite imagery dataset. IEEE Trans Geosci Remote Sens, 57( 1): 574– 586

[20] JiaY Shelhamer E DonahueJ KarayevS LongJ GirshickR ( 2014). Caffe: Convolutional architecture for fast feature embedding. In: Proceedings of the 22nd ACM international conference on multimedia

[21] JiangB YaoX ( 2010). Geospatial analysis and modeling of urban structure and dynamics: an overview. In: Geospatial analysis and modelling of urban structure and dynamics. Dordrecht: Springer, 3– 11

[22] KussulN, LavreniukM, SkakunS, ShelestovA. ( 2017). Deep learning classification of land cover and crop types using remote sensing data. IEEE Geosci Remote Sens Lett, 14( 5): 778– 782

[23] LängkvistM, KiselevA, AlirezaieM, LoutfiA. ( 2016). Classification and segmentation of satellite orthoimagery using convolutional neural networks. Remote Sens (Basel), 8( 4): 329– 329

[24] LecunL, BottouL, BengioY, HaffnerP. ( 1998). Gradient-based learning applied to document recognition. Proc IEEE, 86( 11): 2278– 2324

[25] LiH, JiaY, ZhouY. ( 2018). Urban expansion pattern analysis and planning implementation evaluation based on using fully convolution neural network to extract land range. Neuroquantology, 16( 5): 814– 822

[26] LiuS J, ShiQ, ZhangL P. ( 2021). Few-shot hyperspectral image classification with unknown classes using multitask deep learning. IEEE Trans Geosci Remote Sens, 59( 6): 5085– 5102

[27] LiuS, DingW, LiuC, LiuY, WangY, LiH. ( 2018). Ern: edge loss reinforced semantic segmentation network for remote sensing images. Remote Sens (Basel), 10( 9): 1339– 1362

[28] LiuT, Abd-ElrahmanA. ( 2018). An object-based image analysis method for enhancing classification of land covers using fully convolutional networks and multi-view images of small unmanned aerial system. Remote Sens (Basel), 10( 3): 457

[29] LongJ ShelhamerE DarrellT (2015). Fully convolutional networks for semantic segmentation. In: 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)

[30] MaggioriE, TarabalkaY, CharpiatG, AlliezP. ( 2016). Fully convolutional neural networks for remote sensing image classification. Geosci Remoet Sens Sympos IEEE, 5071– 5074

[31] MartinsV S, KaleitaA L, GelderB K, da SilveiraH L F, AbeC A. ( 2020). Exploring multiscale object-based convolutional neural network (multi-OCNN) for remote sensing image classification at high spatial resolution. ISPRS J Photogramm Remote Sens, 168: 56– 73

[32] MbogaN, GrippaT, GeorganosS, VanhuysseS, SmetsB, DewitteO, WolffE, LennertM. ( 2020). Fully convolutional networks for land cover classification from historical panchromatic aerial photographs. ISPRS J Photogramm Remote Sens, 167: 385– 395

[33] MiddelA, LukasczykJ, ZakrzewskiS, ArnoldM, MaciejewskiR. ( 2019). Urban form and composition of street canyons: a human-centric big data and deep learning approach. Landsc Urban Plan, 183: 122– 132

[34] MohammadimaneshF, SalehiB, MahdianpariM, GillE, MolinierM. ( 2019). A new fully convolutional neural network for semantic segmentation of polarimetric sar imagery in complex land cover ecosystem. ISPRS J Photogramm Remote Sens, 151: 223– 236

[35] NguyenT, HanJ, ParkD C. ( 2013). Satellite image classification using convolutional learning. In: Proceedings of the AIP Conference, Albuquerque

[36] PanG, QiG, WuZ, ZhangD, LiS. ( 2013). Land-use classification using taxi GPS traces. IEEE Trans Intell Transp Syst, 14( 1): 113– 123

[37] PanX, GaoL, MarinoniA, ZhangB, YangF, GambaP. ( 2018). Semantic labeling of high resolution aerial imagery and lidar data with fine segmentation network. Remote Sens (Basel), 10( 5): 743– 767

[38] PerselloC, SteinA. ( 2017). Deep fully convolutional networks for the detection of informal settlements in VHR images. IEEE Geosci Remote Sens Lett, 14( 12): 2325– 2329

[39] PtuchaR, Petroski SuchF, PillaiS, BrocklerF, SinghV, HutkowskiP. ( 2019). Intelligent character recognition using fully convolutional neural networks. Pattern Recognit, 88: 604– 613

[40] QiuC, SchmittM, GeißC, ChenT K, ZhuX X. ( 2020). A framework for large-scale mapping of human settlement extent from Sentinel-2 images via fully convolutional neural networks. ISPRS J Photogramm Remote Sens, 163: 152– 170

[41] ShaoZ, FuH, LiD, AltanO, ChengT. ( 2019). Remote sensing monitoring of multi-scale watersheds impermeability for urban hydrological evaluation. Remote Sens Environ, 232: 111338

[42] SharmaA, LiuX, YangX, ShiD. ( 2017). A patch-based convolutional neural network for remote sensing image classification. Neural Netw, 95: 19– 28

[43] ShiC, PunC M. ( 2019). Adaptive multi-scale deep neural networks with perceptual loss for panchromatic and multispectral images classification. Inform Sciences, 490: 1– 17

[44] ShiQ, LiuM, LiS, LiuX, WangF, ZhangL. ( 2021). A deeply supervised attention metric-based network and an open aerial image dataset for remote sensing change detection. IEEE Trans Geosci Remote Sens, 60: 1– 16

[45] StarkT, WurmM, ZhuX X, TaubenböckH. ( 2020). Satellite-based mapping of urban poverty with transfer-learned slum morphologies. IEEE J Sel Top Appl Earth Obs Remote Sens, 13: 5251– 5263

[46] TanY H, XiongS Z, YanP. ( 2020). Multi-branch convolutional neural network for built-up area extraction from remote sensing image. Neurocomputing, 396: 358– 374

[47] TianY PeiK Jana S RayB (2018). Deeptest: automated testing of deepneural-network-driven autonomous cars. In: Proceedings of the 40th International Conference on Software Engineering

[48] VizzariM, HilalM, SiguraM, AntognelliS, JolyD. ( 2018). Urban-rural-natural gradient analysis with CORINE data: an application to the metropolitan France. Landsc Urban Plan, 171: 18– 29

[49] VolpiM, TuiaD. ( 2017). Dense semantic labeling of subdecimeter resolution images with convolutional neural networks. IEEE Trans Geosci Remote Sens, 55( 2): 881– 893

[50] WagnerR, ThomM, SchweigerR, PalmG, RothermelA. ( 2013). Learning convolutional neural networks from few samples. In The 2013 International Joint Conference on Neural Networks (IJCNN), IEEE: 1– 7

[51] WaldnerF, DiakogiannisF I. ( 2020). Deep learning on edge: extracting field boundaries from satellite images with a convolutional neural network. Remote Sens Environ, 245: 111741

[52] WangQ, GaoJ Y, YuanY. ( 2018). Embedding structured contour and location prior in siamesed fully convolutional networks for road detection. IEEE Trans Intell Transp Syst, 19( 1): 230– 241

[53] WenC, YangL, LiX, PengL, ChiT. ( 2020). Directionally constrained fully convolutional neural network for airborne LiDAR point cloud classification. ISPRS J Photogramm Remote Sens, 162: 50– 62

[54] WengQ. ( 2001). A remote sensing of GIS evaluation of urban expansion and its impact on surface temperature in the Zhujiang Delta, China. Int J Remote Sens, 22( 10): 1999– 2014

[55] WuH Zhang H ZhangJ F XuF J ( 2015). Fast aircraft detection in satellite images based on convolutional neural networks. In: 2015 IEEE International Conference on Image Processing, New York

[56] WurmM, StarkT, ZhuX X, WeigandM, TaubenböckH. ( 2019). Semantic segmentation of slums in satellite images using transfer learning on fully convolutional neural networks. Int J Remote Sens, 150: 59– 69

[57] XiJ Y, NgC N. ( 2007). Spatial and temporal dynamics of urban sprawl along two urban–rural transects: a case study of Guangzhou, China. Landsc Urban Plan, 79( 1): 96– 109

[58] XuY, WuL, XieZ, ChenZ. ( 2018). Building extraction in very high resolution remote sensing imagery using deep learning and guided filters. Remote Sens (Basel), 10( 1): 144– 156

[59] ZeilerM D FergusR ( 2014). Visualizing and Understanding Convolutional Networks. In: European Conference on Computer Vision. Charm: Springer

[60] ZhangC, HarrisonP A, PanX, LiH, SargentI, AtkinsonP M. ( 2020a). Scale sequence joint deep learning (SS-JDL) for land use and land cover classification. Remote Sens Environ, 237: 111593

[61] ZhangC, SargentI, PanX, LiH, GardinerA, HareJ, AtkinsonP M. ( 2018a). An object-based convolutional neural network (OCNN) for urban land use classification. Remote Sens Environ, 216: 57– 70

[62] ZhangC, SargentI, PanX, LiH, GardinerA, HareJ, AtkinsonP M. ( 2019). Joint deep learning for land cover and land use classification. Remote Sens Environ, 221: 173– 187

[63] ZhangC, YueP, TapeteD, ShangguanB, WangM, WuZ. ( 2020b). A multi-level context-guided classification method with object-based convolutional neural network for land cover classification using very high resolution remote sensing images. ITC J, 88: 102086

[64] ZhangD J ZhangJ S PanY Z DuanY M ( 2018b). Fully convolutional neural networks for large scale cropland mapping with historical label dataset. In: IGARSS 2018–2018 IEEE International Geoscience and Remote Sensing Symposium

[65] ZhongY, FeiF, ZhangL. ( 2016). Large patch convolutional neural networks for the scene classification of high spatial resolution imagery. J Appl Remote Sens, 10( 2): 025006

[66] ZhouW, MingD, LvX, ZhouK, BaoH, HongZ. ( 2020). SO–CNN based urban functional zone fine division with VHR remote sensing image. Remote Sens Environ, 236: 111458

[67] ZhuX X, TuiaD, MouL, XiaG S, ZhangL, XuF, FraundorferF. ( 2017). Deep learning in remote sensing: a comprehensive review and list of resources. IEEE Geosci Remote Sens Mag, 5( 4): 8– 36

Acknowledgments

This work is partially supported by Natural Science Foundation of Chongqing in China (No. cstc2020jcyj-jqX0004), the Ministry of education of Humanities and Social Science project (No. 20YJA790016), the National Natural Science Foundation of China (Grant No. 42171298). We thank the patent supporting the method section of the paper (No. 202110750360.1).

Code, Data, and Materials Availability Python3.x and Caffe are required in this paper.

RIGHTS & PERMISSIONS

2022 Higher Education Press