-
摘要: 遥感影像数据与地理信息系统（geographic information system，GIS）矢量数据的配准是遥感与GIS集成的基础。目前遥感影像与矢量数据的配准关键在于遥感影像特征的提取，而现有遥感影像特征提取方法存在特征提取不完整、配准失败和精度不高等问题。由此提出了一种基于Mask R-CNN（region-based convolutional neural network）的遥感影像与矢量数据配准方法，首先，利用Mask R-CNN模型提取影像的道路交叉口作为影像控制点; 然后，依据几何拓扑关系筛选矢量数据道路交叉口作为矢量控制点，再根据遥感影像与矢量数据控制点的欧氏距离确定同名控制点；最后，以同名控制点为基础实现遥感影像与矢量数据的配准。选取上海市矢量数据和高分二号影像数据进行配准实验，实验结果表明, 所提方法鲁棒性强、精度高。
-
关键词:
- 数据配准 /
- 遥感影像 /
- 矢量数据 /
- Mask R-CNN /
- 道路交叉口
Abstract:Objectives The registration of remote sensing image data and GIS (geographic information system) vector data is the basis of the integration of remote sensing and GIS, which is widely used in the fields of map data update, city monitoring, map change detection and so on. At present, the key to the registration of remote sensing images and vector data is the extraction of remote sensing image features. However, the existing remote sensing image feature extraction has problems such as incomplete feature extraction, which leads to registration failure or low accuracy. This paper proposes a registration method for remote sensing images and vector data based on Mask region-based convolutional neural network (Mask R-CNN).Methods Firstly, we select the road intersection as the distinctive feature of the same name in the remote sensing image and vector data, and create a road intersection image data set to train Mask R-CNN model. Secondly, according to the geometric topological relationship, the vector data road intersections are selected as vector control points. And take the intersection control point of the vector data as the center, we use 400 × 400 pixels window to crop the remote sensing image data and input it into the Mask R-CNN model, extract the border of the road intersection in the image. The control points of the same name are determined according to the Euclidean distance between the remote sensing image and the vector data control points, and the control points of the same name are cleaned using the density-based spatial clustering of applications with noise algorithm. Finally, the affine transformation parameters are calculated according to the filtered control points of the same name to realize the registration of remote sensing image and vector data. The registration data of Shanghai vector data and Gaofen-2 image data were selected for registration experiment.Results Experimental results show that the average deviations of the experimental data before were 15.34 m and 1.44 m, before and after registration based on Mask R-CNN.This method can correctly register remote sensing images and vector data, and the proposed method has better application prospects in urban data registration, and has the characteristics of strong robustness and high accuracy.Conclusions The proposed method in this paper can automatically register remote sensing images and vector data from different sources in areas such as mountains, grasslands, and deserts. In the next step, the amount of sample data of remote sensing images will be increased, including types of remote sensing images and types of features.-
Keywords:
- data registration /
- remote sensing image /
- vector data /
- Mask R-CNN /
- road intersection
-
-
![]()
图 5 多车道道路交叉口控制点选取示意图
Figure 5. Schematic Diagram of Control Points Selection for Multi-lane Road Intersections
![]()
图 8 Mask R-CNN与Faster R-CNN训练损失对比图
Figure 8. Comparison of Training Loss Between Mask R-CNN and Faster R-CNN
![]()
图 9 遥感影像道路交叉口提取部分结果图
Figure 9. Extraction Results of Road Intersections in Remote Sensing Images
表 1 遥感影像道路交叉口提取评价结果/%
Table 1 Evaluation Results of Road Intersection Extraction by Remote Sensing Image/%
模型 AP50 AP75 Mask R-CNN 71.3 54.5 Faster R-CNN 60.8 48.5 
下载: 导出CSV
表 2 人工标识同名点配准前后误差距离
Table 2 Errors Distance Before and After Registration of Artificially Identified Points with the Same Name
同名点编号 配准前 基于本文方法配准结果 基于模板匹配配准结果 x/m y/m d/m x'/m y'/m d'/m x″/m y″/m d″/m 1 14.07 13.79 19.70 0.88 0.79 1.18 1.86 1.31 2.28 2 13.31 11.30 17.46 0.33 0.32 0.46 1.85 1.80 2.58 3 14.65 13.73 20.08 2.30 2.29 3.25 2.06 1.65 2.64 4 9.47 8.71 12.86 1.10 1.05 1.52 1.95 1.87 2.70 5 13.49 12.37 18.30 0.22 0.14 0.26 2.28 1.76 2.88 6 12.71 11.49 17.14 2.02 1.99 2.84 2.25 1.79 2.88 7 12.95 11.19 17.11 1.34 1.31 1.87 2.32 2.28 3.25 8 9.58 9.41 13.43 0.63 0.61 0.88 2.43 2.19 3.27 9 11.49 10.36 15.47 0.90 0.82 1.22 2.63 2.04 3.33 10 9.05 8.68 12.54 0.89 0.88 1.25 2.76 2.61 3.80 11 8.91 8.81 12.53 0.99 0.90 1.34 2.97 2.78 4.07 12 10.09 9.06 13.56 0.26 0.25 0.36 3.02 2.82 4.13 13 9.34 8.50 12.63 0.88 0.87 1.24 3.32 3.27 4.66 14 9.72 9.59 13.66 0.60 0.54 0.81 3.82 3.27 5.03 15 9.86 9.48 13.68 2.30 2.21 3.19 4.85 4.83 6.84 平均值/m 11.25 10.43 15.34 1.04 1 1.44 2.69 2.42 3.62 中误差/m 2.06 1.80 2.71 0.68 0.68 0.96 0.82 0.89 1.20 最大值/m 14.65 13.79 20.08 2.30 2.29 3.25 4.85 4.83 6.84
下载: 导出CSV
-
[1] Ceresola S, Fusiello A, Bicego M, et al. Automatic Updating of Urban Vector Maps[C]// Image Analysis and Processing ICIAP, Cagliari, Italy, 2005
[2] Durieux L, Lagabrielle E, Nelson A. A Method for Monitoring Building Construction in Urban Sprawl Areas Using Object-Based Analysis of Spot 5 Images and Existing GIS Data[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2008, 63(4): 399-408 doi: 10.1016/j.isprsjprs.2008.01.005
[3] 李全, 李霖, 赵曦. 基于Landsat TM影像的城市变化检测研究[J]. 武汉大学学报· 信息科学版, 2005, 30(4): 351-354 http://ch.whu.edu.cn/article/id/2158 Li Quan, Li Lin, Zhao Xi. Urban Change Detection Using Landsat TM Imagery[J]. Geomatics and Information Science of Wuhan University, 2005, 30 (4): 351-354 http://ch.whu.edu.cn/article/id/2158
[4] Champion N. 2D Building Change Detection from High Resolution Aerial Images and Correlation Digital Surface Models[J]. Remote Sensing and Spatial Information Sciences, 2007, 36(3): 197–202
[5] 佃袁勇, 方圣辉, 姚崇怀. 一种面向地理对象的遥感影像变化检测方法[J]. 武汉大学学报·信息科学版, 2014, 39(8): 906-912 doi: 10.13203/j.whugis20130053 Dian Yuanyong, Fang Shenghui, Yao Chonghuai. The Geographic Object-Based Method for Change Detection with Remote Sensing Imagery[J]. Geomatics and Information Science of Wuhan University, 2014, 39(8): 906-912 doi: 10.13203/j.whugis20130053
[6] 张剑清, 朱丽娜, 潘励. 基于遥感影像和矢量数据的水系变化检测[J]. 武汉大学学报·信息科学版, 2007, 32(8): 663-666 http://ch.whu.edu.cn/article/id/1951 Zhang Jianqing, Zhu Lina, Pan Li. River Change Detection Based on Remote Sensing Imagery and Vector Data[J]. Geomatics and Information Science of Wuhan University, 2007, 32(8): 663-666 http://ch.whu.edu.cn/article/id/1951
[7] Guo Z, Du S H, Zhao W Z, et al. A Graph-Based Approach for the Co-registration Refinement of Very-High-Resolution Imagery and Digital Line Graphic Data[J]. International Journal of Remote Sensing, 2016, 37(17): 4015-4034 doi: 10.1080/01431161.2016.1207259
[8] 刘志青, 郭海涛, 陈小卫, 等. 一种航空影像与矢量数据配准的新方法[J]. 测绘科学, 2013, 38(5): 157-161 https://www.cnki.com.cn/Article/CJFDTOTAL-CHKD201305052.htm Liu Zhiqing, Guo Haitao, Chen Xiaowei, et al. Registration Method Between Aerial Image and Vector Data Based on 2D Multi-scale Template[J]. Science of Surveying and Mapping, 2013, 38(5): 157-161 https://www.cnki.com.cn/Article/CJFDTOTAL-CHKD201305052.htm
[9] 赵珍珍, 燕琴, 刘正军. 高分遥感影像与矢量数据结合的变化检测方法[J]. 测绘科学, 2015, 40(6): 120-124 https://www.cnki.com.cn/Article/CJFDTOTAL-CHKD201506025.htm Zhao Zhenzhen, Yan Qin, Liu Zhengjun. Research of Change Detection Using High-Resolution Remote Sensing Images and Vector Data Oriented to Geographic National Conditions Monitoring[J]. Science of Surveying and Mapping, 2015, 40(6): 120-124 https://www.cnki.com.cn/Article/CJFDTOTAL-CHKD201506025.htm
[10] 王洪华, 郭建星. 遥感图像融合技术及其在更新GIS数据库中的应用[J]. 测绘通报, 2003(2): 11-13 https://www.cnki.com.cn/Article/CJFDTOTAL-CHTB200302004.htm Wang Honghua, Guo Jianxing. RS Image Merging and Its Use in GIS Database Updating[J]. Bulletin of Surveying and Mapping, 2003(2): 11-13 https://www.cnki.com.cn/Article/CJFDTOTAL-CHTB200302004.htm
[11] 张剑清, 董明, 张宏伟. TM影像与GIS矢量数据的自动配准[J]. 武汉大学学报·信息科学版, 2005, 30 (11): 950-954 http://ch.whu.edu.cn/article/id/2312 Zhang Jianqing, Dong Ming, Zhang Hongwei. Automatically Registration of TM Image and GIS Vector Data[J]. Geomatics and Information Science of Wuhan University, 2005, 30(11): 950-954 http://ch.whu.edu.cn/article/id/2312
[12] 张晓东, 李德仁, 龚健雅, 等. 一种基于面特征的遥感影像与GIS数据配准方法[J]. 遥感学报, 2006, 10(3): 373-380 https://www.cnki.com.cn/Article/CJFDTOTAL-YGXB200603013.htm Zhang Xiaodong, Li Deren, Gong Jianya, et al. A Matching Method of Remote Sensing Image and GIS Data Based on Area Feature[J]. Journal of Remote Sensing, 2006, 10(3): 373-380 https://www.cnki.com.cn/Article/CJFDTOTAL-YGXB200603013.htm
[13] 张江水, 李传广, 郭海涛. 基于动态规划和Hough变换的遥感影像与GIS矢量数据匹配方法[J]. 测绘工程, 2011, 20 (5): 9-12 https://www.cnki.com.cn/Article/CJFDTOTAL-CHGC201105004.htm Zhang Jiangshui, Li Chuanguang, Guo Haitao. A Matching Method of Remote Sensing Image and GIS Vector Data Based on Dynamic Programming and Hough Transform[J]. Engineering of Surveying and Mapping, 2011, 20(5): 9-12 https://www.cnki.com.cn/Article/CJFDTOTAL-CHGC201105004.htm
[14] Drewniok C, Rohr K. Automatic Exterior Orienta‐ tion of Aerial Images in Urban Environments[C]//2008 ISPRS Congress, Beijing, China, 2008
[15] Vasileisky A, Zhukov B, Berger M. Automated Image Co-registration Based on Linear Feature Recognition[C]//The 2nd Conference Fusion of Earth Data, Paris, France, 1998
[16] Habbecke M, Kobbelt L. Automatic Registration of Oblique Aerial Images with Cadastral Maps[M]. Heidelberg: Springer, 2010
[17] Pehani P, Čotar K, Marsetič A, et al. Automatic Geometric Processing for very High Resolution Optical Satellite Data Based on Vector Roads and Orthophotos[J]. Remote Sensing, 2016, 8(4): 343 doi: 10.3390/rs8040343
[18] Schickler W. Feature Matching for Outer Orientation of Single Images Using 3D Wireframe Contro Points[C]//International Archives of Photogrammetry and Remote Sensing, Graz, Austria, 1990
[19] He K M, Gkioxari G, Dollár P, et al. Mask RCNN[C]// IEEE International Conference on Computer Vision, Venice, Italy, 2017
[20] Ester M, Kriegel H P, Sander J, et al. A DensityBased Algorithm for Discovering Clusters in Large Spatial Databases with Noise[C]//The 2nd Conference on Knowledge Discovery and Data Mining, Portland, Oregon, 1996
[21] Lin T Y, Dollár P, Girshick R, et al. Feature Pyramid Networks for Object Detection[C]// IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, HI, USA, 2017
[22] GB 50647-2011. Code for Planning of Intersections on Urban Roads[S]. Ministry of Housing and Urban-Rural Development of the People's Republic of China, 2011 GB 50647-2011. 城市道路交叉口规划规范[S]. 中华人民共和国住房和城乡建设部, 2011
[23] 王晓静, 王铁军, 许高升. 一种基于仿射变换模型的图像自动对准方法[J]. 战术导弹技术, 2008(5): 73-77 https://www.cnki.com.cn/Article/CJFDTOTAL-ZSDD200805015.htm Wang Xiaojing, Wang Tiejun, Xu Gaosheng. A New Image Registration Approach Based on Affine Transformation Model[J]. Tactical Missile Technology, 2008(5): 73-77 https://www.cnki.com.cn/Article/CJFDTOTAL-ZSDD200805015.htm
[24] He K M, Zhang X Y, Ren S Q, et al. Deep Residual Learning for Image Recognition [C]// IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, NV, USA, 2016
[25] Guo X R, Zhang W Y, Ma G B. Automatic Urban Remote Sensing Images Registration Based on Road Networks[C]//2009 Joint Urban Remote Sensing Event, Shanghai, China, 2009
[26] Ren S Q, He K M, Girshick R, et al. Faster RCNN: Towards Real-Time Object Detection with Region Proposal Networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(6): 1137-1149 doi: 10.1109/TPAMI.2016.2577031
-
期刊类型引用(4)
1. 刘晓云，郭春喜，靳鑫洋，蒋涛. 2020年珠峰高程测量与确定流程解析. 大地测量与地球动力学. 2024(02): 111-115+127 . 
百度学术
2. 张建华，张庆涛，张涛，郑文科，陈小英. 顾及水准起算重力异常差异的珠峰地区垂直形变分析. 测绘科学. 2023(06): 1-8 . 百度学术
3. 郭春喜，靳鑫洋，蒋涛，王斌，刘晓云. 2020与2005珠峰测量与高程确定异同. 测绘科学. 2023(07): 10-15 . 百度学术
4. 党亚民，蒋涛，杨元喜，孙和平，姜卫平，朱建军，薛树强，张小红，蔚保国，罗志才，李星星，肖云，章传银，张宝成，李子申，冯伟，任夏，王虎. 中国大地测量研究进展（2019—2023）. 测绘学报. 2023(09): 1419-1436 . 百度学术
其他类型引用(0)
计量
- 文章访问数: 1185
- HTML全文浏览量: 333
- PDF下载量: 144
- 被引次数: 4
