基于开源数据和条件生成对抗网络的地形重建方法

doi:10.12082/dqxxkx.2023.220701

地球信息科学学报 ›› 2023, Vol. 25 ›› Issue (2): 252-264.doi: 10.12082/dqxxkx.2023.220701

• 地球信息科学理论与方法 • 上一篇下一篇

基于开源数据和条件生成对抗网络的地形重建方法

陈凯¹^,²(), 雷少华³, 代文¹^,^*(), 王春², 刘爱利¹, 李敏²

1.南京信息工程大学地理科学学院，南京 210044
2.实景地理环境安徽省重点实验室，滁州 239000
3.南京水利科学研究院水文水资源与水利工程科学国家重点实验室，南京 210029

收稿日期:2022-09-19 修回日期:2022-10-31 出版日期:2023-02-25 发布日期:2023-04-19
通讯作者: *代文（1995— ），男，贵州毕节人，博士，讲师，研究方向为实景三维建模与数字地形分析。 E-mail: wen.dai@nuist.edu.cn
作者简介:陈凯（1999— ），男，甘肃天水人，硕士生，研究方向为实景三维建模与数字地形分析。E-mail: 20211210002@nuist.edu.cn
基金资助:
江苏省高等学校自然科学研究项目(22KJB170016);国家自然科学基金项目(42101384);江苏省自然科学基金青年基金项目(BK20210043);安徽高校省级自然科学研究重大项目(KJ2021ZD0130);江西省水利厅重大科技项目(202124ZDKT29);“实景地理环境安徽省重点实验室”开放课题资助(2022PGE013)

Terrain Rebuilding Method based on Open Source Data and Conditional Generative Adversarial Networks

CHEN Kai¹^,²(), LEI Shaohua³, DAI Wen¹^,^*(), WANG Chun², LIU Aili¹, LI Min²

1. School of Geographical Sciences, Nanjing University of Information Science and Technology, Nanjing 210044, China
2. Key Laboratory of Physical Geographic Information in Anhui Province, Chuzhou 239000, China
3. State Key Laboratory of Hydrology-Water Resources and Hydraulic Engineering, Nanjing Hydraulic Research Institute, Nanjing 210029, China

Received:2022-09-19 Revised:2022-10-31 Online:2023-02-25 Published:2023-04-19
Contact: DAI Wen
Supported by:
The Natural Science Foundation of the Jiangsu Higher Education Institutions of China(22KJB170016);The National Natural Science Foundation of China, No.42101384(42101384);The Natural Science Foundation of Jiangsu Province(BK20210043);The Major Project of Natural Science Research of Anhui Provincial Department of Education(KJ2021ZD0130);The Water Conservancy Science and Technology Project of Jiangxi Province(202124ZDKT29);The Foundation of Anhui Province Key Laboratory of Physical Geographic Environment(2022PGE013)

摘要/Abstract

摘要：

如何使用少量的地形特征复原地形地貌一直为地学领域的难题。本文使用开源数据集提取地形特征要素，使用地形特征要素作为约束条件构建了用于生成DEM的条件生成对抗网络(Conditional Generative Adversarial Networks, CGAN)，设计了基于开源DEM、开源DEM与遥感影像组合、以及5m高精度DEM提取地形特征要素生成DEM的对比实验，并对结果进行视觉效果、相关性分析以及地形因子的对比与评价。结果表明：① 在视觉效果上，3种不同方式生成的DEM在视觉效果上均十分逼近原始5 m DEM，都远好于传统插值方法生成DEM，基于开源12.5m DEM提取要素和1m遥感影像的重建效果最接近于原始5 m DEM；② 在相关性上，三种不同方式生成的DEM与原始5m DEM相关性均能达到0.75以上，组合开源数据提取要素重建DEM与原始5 m DEM相关性可达到0.85以上；③ 在地形因子方面，基于开源12.5 m DEM和1 m遥感影像提取要素重建DEM的坡度和坡向的分布趋势与原始5 m DEM最为一致。本文为高精度DEM建模提供了新的思路，在高精度DEM难以获取的区域，可以利用开源数据集和条件生成对抗网络进行高精度地形建模，从而进行地学分析和地理模拟等。

关键词: 开源数据, 遥感影像, 条件生成对抗网络, 数字高程模型, 地形特征要素, 地形重建, 地形骨架, 黄土高原

Abstract:

How to use a small number of topographic features to restore the topography has been a difficult problem in the field of geology. In this paper, we extract topographic features from open source datasets, and construct Conditional Generative Adversarial Networks (CGAN) for DEM generation using topographic features as constraints, a comparative experiment was designed based on the combination of open-source DEM, open-source DEM and remote sensing image, as well as the generation of DEM by extracting topographic features from the high-precision DEM with a resolution of 5 m, the results were compared and evaluated by visual effect, correlation analysis and topographic factors. The results show that: (1) in the visual effect, the DEM generated by three different methods are very close to the original DEM with a resolution of 5m, which is much better than the traditional interpolation method, (2) the correlation between DEM generated by three different methods and the original DEM with a resolution of 5m is more than 0.75, and the result of reconstruction based on dem with a resolution of 5 m extracted from open source and remote sensing image with a resolution of 1m is closest to that of the original DEM with a resolution of 5m, the correlation between DEM and original 5m DEM can reach more than 0.85. (3) in the aspect of terrain factor, based on dem with a resolution of 5 m and remote sensing image with a resolution of 1m, the distribution trend of slope and aspect of reconstructed DEM is most consistent with the original DEM with a resolution of 5 m. This paper provides a new idea for high-precision DEM modeling. In the areas where high-precision DEM is difficult to obtain, high-precision terrain modeling can be carried out by using open source data sets and Conditional Generative Adversarial Networks, so as to conduct geoscience analysis and geographical simulation.

Key words: open source data, remote sensing images, conditional generative adversarial networks, Digital Elevation Model, topographic features, topographic rebuilding, terrain skeleton, Loess Plateau

陈凯, 雷少华, 代文, 王春, 刘爱利, 李敏. 基于开源数据和条件生成对抗网络的地形重建方法[J]. 地球信息科学学报, 2023, 25(2): 252-264.DOI:10.12082/dqxxkx.2023.220701

CHEN Kai, LEI Shaohua, DAI Wen, WANG Chun, LIU Aili, LI Min. Terrain Rebuilding Method based on Open Source Data and Conditional Generative Adversarial Networks[J]. Journal of Geo-information Science, 2023, 25(2): 252-264.DOI:10.12082/dqxxkx.2023.220701

图/表 12

表1

图1

图2

图3

图4

图5

图6

图7

图8

图9

图10

表2

参考文献 33

[1]	徐亚莉, 罗明良, 梁倍瑜, 等. DEM空间插值方法对土壤侵蚀模拟的影响研究——以USPED分析干热河谷典型冲沟为例[J]. 地理科学进展, 2016, 35(7):870-877. doi: 10.18306/dlkxjz.2016.07.008
	[ Xu Y L, Luo M L, Liang B Y, et al. Effects of different DEM spatial interpolation methods on soil erosion simulation: A case study of a typical gully of dry-hot valley based on USPED[J]. Progress in Geography, 2016, 35(7):870-877. ]
[2]	陈传法, 岳天祥. 基于条件随机模拟的DEM误差分析——以董志塬水土流失等级划分为例[J]. 地球信息科学学报, 2009, 11(5):5572-5576.
	[ Chen C F, Yue T X. DEM error analysis based on conditional stochastic simulation[J]. Journal of Geo-information Science, 2009, 11(5):5572-5576. ]
[3]	兰进京, 于海洋, 陈琳, 等. 机载LiDAR DEM在流域水文分析与模拟中的尺度效应[J]. 测绘通报, 2020(4):40-46.
	[ Lan J J, Yu H Y, Chen L, et al. Scale effect of airborne LiDAR DEM in watershed hydrological analysis and simulation[J]. Bulletin of Surveying and Mapping, 2020(4):40-46. ] DOI:10.13474/j.cnki.11-2246.2020.0109. doi: 10.13474/j.cnki.11-2246.2020.0109
[4]	Dai W, Qian W, Liu A L, et al. Monitoring and modeling sediment transport in space in small loess catchments using UAV-SfM photogrammetry[J]. CATENA, 2022, 214:106244. DOI:10.1016/j.catena.2022.106244. doi: 10.1016/j.catena.2022.106244
[5]	王金鑫, 秦子龙, 赵光成, 等. 基于球体网格与DEM的流域地形特征尺度效应分析——以长江流域为例[J]. 应用基础与工程科学学报, 2022, 30(5):1109-1120.
	Qin Z L, Zhao G C, et al. Scale Effect analysis of basin topographic features based on spherical grid and DEM: taking the Yangtze River Basin as an example[J]. Journal of Basic Science and Engineering, 2022, 30(5):1109-1120. ] DOI:10.16058/j.issn.1005-0930.2022.05.005 doi: 10.16058/j.issn.1005-0930.2022.05.005
[6]	Dai W, Yang X, Na J M, et al. Effects of DEM resolution on the accuracy of gully maps in loess hilly areas[J]. CATENA, 2019, 177:114-125. DOI:10.1016/j.catena.2019.02.010 doi: 10.1016/j.catena.2019.02.010
[7]	毛玉辉, 宋众艳, 岳忠奇, 等. 灰度图在地形建模中的应用[J]. 火力与指挥控制, 2019, 44(11):160-164.
	[ Mao Y H, Song Z Y, Yue Z Q, et al. Application of grey scale image in terrain modeling[J]. Fire Control ＆ Command Control, 2019, 44(11):160-164. ] DOI:10.3969/j.issn.1002-06 40.2019.11.032 doi: 10.3969/j.issn.1002-06 40.2019.11.032
[8]	徐建华, 顾浩, 曾艳阳, 等. 自适应屏幕分辨率的海底地形建模研究[J]. 计算机仿真, 2022, 39(4):142-145.
	[ Xu J H, Gu H, Zeng Y Y, et al. Research on seabed terrain modeling with adaptive screen resolution[J]. Computer Simulation, 2022, 39(4):142-145. ]
[9]	代文, 那嘉明, 杨昕, 等. 基于DEM光照晕渲模拟的梯田自动提取方法[J]. 地球信息科学学报, 2017, 19(6):754-762. doi: 10.3724/SP.J.1047.2017.00754
	[ Dai W, Na J M, Yang X, et al. An automatic retrieval method for artificial terrace based on illumination model of DEM shading[J]. Journal of Geo-information Science, 2017, 19(6):754-762. ]
[10]	毕瑞, 甘淑, 李绕波, 等. 面向东川复杂山地泥石流沟谷三维地形建模及特征分析的无人机遥感探测应用研究[J]. 中国地质灾害与防治学报, 2021, 32(3):91-100.
	[ Bi R, Gan S, Li R B, et al. Application research of unmanned aerial vehicle remote sensing detection for 3D terrain modeling and feature analysis of debris flow gullies in complex mountainous area of Dongchuan[J]. The Chinese Journal of Geological Hazard and Control, 2021, 32(3):91-100. ] DOI:10.16031/j.cnki.issn.1003-8035.2021.03-12 doi: 10.16031/j.cnki.issn.1003-8035.2021.03-12
[11]	James M R, Robson S, D'oleire-Oltmanns S, et al. Optimising UAV topographic surveys processed with structure-from-motion: Ground control quality, quantity and bundle adjustment[J]. Geomorphology, 2017, 280:51-66. DOI:10.1016/j.geomorph.2016.11.021 doi: 10.1016/j.geomorph.2016.11.021
[12]	李清, 侯永军, 沈春林. 数字地形数据的二维三次卷积插值[J]. 南京航空航天大学学报, 1997, 29(4):378-384.
	[ Li Q, Hou Y J, Shen C L. Two dimensional cubic convolution interpolation for digital map data[J]. Transactions of Nanjing University of Aeronautics ＆ Astronautics, 1997, 29(4): 378-384. ]
[13]	刘世彬, 刘兴彦. 最小二乘内插数学模型生成虚拟地形的可视化研究[J]. 计算机工程与科学, 2008, 30(9):29-31.
	[ Liu S B, Liu X Y. Research on the visualization of the virtual terrain generated by the least-squares interpolation mathematical model[J]. Computer Engineering ＆ Science, 2008, 30(9):29-31. ]
[14]	刘世彬, 刘兴彦. 多面函数内插数学模型生成虚拟地形的可视化研究[J]. 火力与指挥控制, 2009, 34(9):28-30.
	[ Liu S B, Liu X Y. Visualization for virtual terrain generated by multi-face function interpolated mathematic model[J]. Fire Control and Command Control, 2009, 34(9): 28-30. ]
[15]	王耀革, 王鑫, 朱长青. 基于双线性内插规则格网DEM地形误差模型[J]. 测绘科学技术学报, 2007, 24(6):419-421.
	[ Wang Y G, Wang X, Zhu C Q. A terrain error model of grid DEM based on bilinear polynomial[J]. Journal of Zhengzhou Institute of Surveying and Mapping, 2007, 24(6):419-421. ]
[16]	Krištof P, Beneš B, Křivánek J, et al. Hydraulic erosion using smoothed particle hydrodynamics[J]. Computer Grap- hics Forum, 2009, 28(2):219-228. DOI:10.1111/j.146 7-8659.2009.01361.x doi: 10.1111/j.146 7-8659.2009.01361.x
[17]	Cordonnier G, Galin E, Gain J, et al. Authoring landscapes by combining ecosystem and terrain erosion simulation[J]. ACM Transactions on Graphics, 2017, 36(4):1-12. DOI:10.1145/3072959.3073667 doi: 10.1145/3072959.3073667
[18]	Cordonnier G, Cani M P, Benes B, et al. Sculpting Mountains: Interactive terrain modeling based on subsurface geology[J]. IEEE Transactions on Visualization and Computer Graphics, 2018, 24(5):1756-1769. DOI:10.1109/TV CG.2017.2689022 doi: 10.1109/TV CG.2017.2689022
[19]	Li S J, Hu G H, Cheng X H, et al. Integrating topographic knowledge into deep learning for the void-filling of digital elevation models[J]. Remote Sensing of Environment, 2022, 269:112818. DOI:10.1016/j.rse.2021.112818 doi: 10.1016/j.rse.2021.112818
[20]	Li S J, Li K, Xiong L Y, et al. Generating terrain data for geomorphological analysis by integrating topographical features and conditional generative adversarial networks[J]. Remote Sensing, 2022, 14(5):1166. DOI:10.3390/rs14051166 doi: 10.3390/rs14051166
[21]	江西林, 王少荣, 李文玉, 等. 以特征线和高程范围为约束的DEM-cGAN框架[J]. 计算机辅助设计与图形学学报, 2021, 33(8):1191-1201.
	[ Jiang X L, Wang S R, Li W Y, et al. DEM-cGAN framework constrained by feature lines and elevation range[J]. Journal of Computer-Aided Design ＆ Computer Graphics, 2021, 33(8):1191-1201. ]
[22]	Isola P, Zhu J Y, Zhou T H, et al. Image-to-image translation with conditional adversarial networks[C]// 2017 IEEE Conference on Computer Vision and Pattern Recognition. IEEE,: 5967-5976. DOI:10.1109/CVPR.2017.632 doi: 10.1109/CVPR.2017.632
[23]	Ronneberger O, Fischer P, Brox T. U-net: Convolutional networks for biomedical image segmentation[M]// Lecture Notes in Computer Science. Cham: Springer International Publishing, 2015:234-241. DOI:10.1007/978-3-31 9-24574-4_28
[24]	唐贤伦, 杜一铭, 刘雨微, 等. 基于条件深度卷积生成对抗网络的图像识别方法[J]. 自动化学报, 2018, 44(5):855-864.
	[ Tang X L, Du Y M, Liu Y W, et al. Image recognition with conditional deep convolutional generative adversarial networks[J]. Acta Automatica Sinica, 2018, 44(5):855-864. ] DOI:10.16383/j.aas.2018.c170470 doi: 10.16383/j.aas.2018.c170470
[25]	张勇. 黄土高原地面坡谱研究[D]. 西安: 西北大学, 2003.
	[ Zhang Y. A research on slope spectrum of the loess plateau[D]. Xi'an: Northwest University, 2003. ]
[26]	代文, 汤国安, 胡光辉, 等. 基于无人机摄影测量的地形变化检测方法与小流域输沙模型研究[J]. 地理科学进展, 2021, 40(9):1570-1580. doi: 10.18306/dlkxjz.2021.09.011
	[ Dai W, Tang G A, Hu G H, et al. Modelling sediment transport in space in a watershed based on topographic change detection by UAV survey[J]. Progress in Geography, 2021, 40(9):1570-1580. ] doi: 10.18306/dlkxjz.2021.09.011
[27]	王静, 张军, 代文, 等. 黄土高原县域贫困空间格局及地形影响因素研究[J]. 地域研究与开发, 2020, 39(3):30-35,52.
	[ Wang J, Zhang J, Dai W, et al. Spatial distribution and topographic influencing factors of poverty-stricken counties in the loess plateau of China[J]. Areal Research and Development, 2020, 39(3):30-35,52. ]
[28]	晏实江, 汤国安, 李发源, 等. 利用DEM边缘检测进行黄土地貌沟沿线自动提取[J]. 武汉大学学报·信息科学版, 2011, 36(3):363-367.
	[ Yan S J, Tang G A, Li F Y, et al. An edge detection based method for extraction of loess shoulder-line from grid DEM[J]. Geomatics and Information Science of Wuhan University, 2011, 36(3):363-367. ] DOI:10.13203/j.whugis2011.03.015 doi: 10.13203/j.whugis2011.03.015
[29]	李可. 基于深度学习的虚拟地形生成方法研究[D]. 南京: 南京师范大学, 2019.
	[ Li K. Virtual terrain generation based on deep learning[D]. Nanjing: Nanjing Normal University, 2019. ]
[30]	朱红春, 汤国安, 张友顺, 等. 基于DEM提取黄土丘陵区沟沿线[J]. 水土保持通报, 2003, 23(5):43-45,61.
	[ Zhu H C, Tang G A, Zhang Y S, et al. Thalweg in loess hill area based on DEM[J]. Bulletin of Soil and Water Conservation, 2003, 23(5):43-45,61. ] DOI:10.13961/j.cn ki.stbctb. 2003.05.011 doi: 10.13961/j.cn ki.stbctb. 2003.05.011
[31]	Zhou Y, Yang C Q, Li F, et al. Spatial distribution and influencing factors of Surface Nibble Degree index in the severe gully erosion region of China's Loess Plateau[J]. Journal of Geographical Sciences, 2021, 31(11):1575-1597. DOI:10.1007/s11442-021-1912-2 doi: 10.1007/s11442-021-1912-2
[32]	Liu K, Ding H, Tang G A, et al. Large-scale mapping of gully-affected areas: An approach integrating Google Earth images and terrain skeleton information[J]. Geomorphology, 2018, 314:13-26. DOI:10.1016/j.geomorp h.201 8.04.011 doi: 10.1016/j.geomorp h.201 8.04.011
[33]	张磊, 汤国安, 李发源, 等. 黄土地貌沟沿线研究综述[J]. 地理与地理信息科学, 2012, 28(6):44-48.
	[ Zhang L, Tang G A, Li F Y, et al. A review on research of loess shoulder-line[J]. Geography and Geo-information Science, 2012, 28(6):44-48. ]

序号	实验名称	实验目的
1	基于5 m DEM地形特征要素的地形重建生成实验	作为对照组，评估网络模型精度
2	基于12.5 m DEM地形特征要素的地形重建实验	探究单一低分辨率开源DEM数据地形特征要素生成高分辨率地形的效果
3	基于1 m遥感影像+12.5 m DEM组合开源数据地形重建实验	1 m遥感影像可以提高沟沿线精度，因此将其组合；探究组合开源数据地形特征要素生成高分辨率地形的效果

		① 基于5 m地形特征要素的重建	② 基于12.5 m地形特征要素的重建	③ 基于组合开源地形特征要素的重建
视觉效果		③ > ① > ②
高程相关性		0.87~0.96	0.77~0.92	0.83~0.96
地形因子	坡度分布	③ > ① > ②
地形因子	坡向分布	③ > ①> ②

基于开源数据和条件生成对抗网络的地形重建方法

Terrain Rebuilding Method based on Open Source Data and Conditional Generative Adversarial Networks

RichHTML

PDF (PC)

赞

可视化

摘要/Abstract

引用本文

使用本文

图/表 12

参考文献 33

相关文章 15

Metrics

本文评价

推荐阅读 0

[1]	涂平, 林偲蔚, 周千千, 谢静, 范玉贵, 陈楠, 林伟斌, 郭连鹤. 顾及地形影响的火星地表天文辐射时空特征分析[J]. 地球信息科学学报, 2023, 25(8): 1611-1624.
[2]	董勇, 周亮, 高鸿, 王宝. 黄土高原地形破碎化探测及空间异质性分析[J]. 地球信息科学学报, 2023, 25(8): 1625-1636.
[3]	张俊瑶, 杨晓梅, 王志华, 杨海坤, 张博淳, 万庆, 雷梅. 绿色发展理念视角下内蒙古煤矿区格局演变分析[J]. 地球信息科学学报, 2023, 25(8): 1655-1668.
[4]	林娜, 何静, 王斌, 唐菲菲, 周俊宇, 郭江. 结合植被光谱特征与Sep-UNet的城市植被信息智能提取方法[J]. 地球信息科学学报, 2023, 25(8): 1717-1729.
[5]	吴敏, 张明达, 李盼盼, 张勇健. 面向多源遥感影像数据的溯源模型研究[J]. 地球信息科学学报, 2023, 25(7): 1325-1335.
[6]	李琳叶, 李艳艳, 陈传法, 刘妍, 刘雅婷, 刘盼盼. 林区数字高程模型修正方法：顾及高程自相关的后向传播神经网络模型[J]. 地球信息科学学报, 2023, 25(5): 935-952.
[7]	衡雪彪, 许捍卫, 唐璐, 汤恒, 许怡蕾. 基于改进全卷积神经网络模型的土地覆盖分类方法研究[J]. 地球信息科学学报, 2023, 25(3): 495-509.
[8]	高鹏飞, 曹雪峰, 李科, 游雄. 融合多元稀疏特征与阶层深度特征的遥感影像目标检测[J]. 地球信息科学学报, 2023, 25(3): 638-653.
[9]	黄帅元, 董有福, 李海鹏. 黄土高原区SRTM1 DEM高程误差校正模型构建及对比分析[J]. 地球信息科学学报, 2023, 25(3): 669-681.
[10]	贝祎轩, 陈传法, 王鑫, 孙延宁, 何青鑫, 李坤禹. 机载LiDAR点云密度和插值方法对DEM及地表粗糙度精度影响分析[J]. 地球信息科学学报, 2023, 25(2): 265-276.
[11]	饶子昱, 卢俊, 郭海涛, 余东行, 侯青峰. 利用视角转换的跨视角影像匹配方法[J]. 地球信息科学学报, 2023, 25(2): 368-379.
[12]	王龙号, 蓝朝桢, 姚富山, 侯慧太, 武蓓蓓. 多源遥感影像深度特征融合匹配算法[J]. 地球信息科学学报, 2023, 25(2): 380-395.
[13]	张建兵, 严泽枭, 马淑芳. 用于遥感影像建筑物变化检测的多尺度交叉对偶注意力网络[J]. 地球信息科学学报, 2023, 25(12): 2487-2500.
[14]	谭敏, 林惠晶, 郝明. 先验知识引导深度学习的耕地范围建筑物和大棚房监测方法[J]. 地球信息科学学报, 2023, 25(11): 2293-2302.
[15]	蒋伟杰, 张春菊, 徐兵, 罗晨晨, 周晗, 周康. AED-Net：滑坡灾害遥感影像语义分割模型[J]. 地球信息科学学报, 2023, 25(10): 2012-2025.