A method for restoring road centerlines based on semantic recognition of skeleton point topology

doi:10.11947/j.AGCS.2026.20260164

Abstract

Abstract:

Fully extracting road centerlines and reconstructing their topology is a crucial step in creating high-quality vector road networks from high-resolution remote sensing images. This process is invaluable for updating geographic information databases and supporting the development of smart cities. However, due to complex surface environments and image occlusions, existing automatic extraction methods often struggle to balance geometric accuracy with topological completeness. Common issues such as broken centerlines, distorted intersections, and missing connectivity significantly limit the reliability and usefulness of road network data derived from remote sensing images. To address this problem, this paper proposes a method for repairing and vectorizing road centerlines by combining skeleton morphology analysis with topological semantic recognition. First, it constructs basic road skeleton units using morphological thinning. By introducing topological invariant analysis, it builds multi-dimensional feature descriptors for skeleton points, enabling precise topological semantic identification of discrete points, line endpoints, points along lines, and intersections. Next, for complex road intersections, it designs a local structure decomposition strategy based on neighborhood morphology constraints. By establishing connectivity repair criteria for breakpoint pairs, it achieves topological reconstruction and geometric regularization of fractured skeletons. Finally, through smooth tracing, it generates a vector road network with complete topological relationships. We validated our method using the DeepGlobe and Massachusetts datasets. The results demonstrate that, compared to the original ME-Net extraction outcomes, the centerline node offset distance and the average road network offset distance decreased by 7.69% and 8.00%, respectively, after applying our repair method. Additionally, node completeness increased to 96.56%, and the average processing time was reduced by 4.63%. Tests have demonstrated that this method effectively overcomes issues such as topological breaks caused by road obstructions and distortions at intersections. It significantly improves the geometric accuracy and topological connectivity of vector road networks, providing reliable theoretical and technical support for automated road network production.

Key words: topology repair, road centerline, skeleton extraction, road vectorization, vector map update

CLC Number:

P208

Ke Zhang, Wenyue Guo, Xin Chen, Liuxin Ren, Junming Chen, Zheng Zhang. A method for restoring road centerlines based on semantic recognition of skeleton point topology[J]. Acta Geodaetica et Cartographica Sinica, 2026, 55(8): 1496-1510.

Add to citation manager EndNote|Reference Manager|ProCite|BibTeX|RefWorks

URL: http://xb.chinasmp.com/EN/10.11947/j.AGCS.2026.20260164

http://xb.chinasmp.com/EN/Y2026/V55/I8/1496

Figures/Tables 13

Fig. 1

Fig. 2

Fig. 3

Fig. 4

Fig. 5

Fig. 6

Fig. 7

Fig. 8

Fig. 9

Tab. 1

Fig. 10

Fig. 11

Fig. 12

References 38

[1]	孙群, 温伯威, 陈欣. 多源地理空间数据一致性处理研究进展[J]. 测绘学报, 2022, 51(7): 1561-1574. DOI: . doi: 10.11947/j.AGCS.2022.20220151
	Sun Qun, Wen Bowei, Chen Xin. Research on consistency processing of multi-source geospatial data[J]. Acta Geodaetica et Cartographica Sinica, 2022, 51(7): 1561-1574. DOI: . doi: 10.11947/j.AGCS.2022.20220151
[2]	孙群, 任福, 龙毅, 等. 中国地图制图学与地理信息工程进展报告(2019—2023)[J]. 测绘学报, 2024, 53(3): 399-412. DOI: . doi: 10.11947/j.AGCS.2024.20230562
	Sun Qun, Ren Fu, Long Yi, et al. The progress and trend of cartography and geographic information engineering in China (2019—2023)[J]. Acta Geodaetica et Cartographica Sinica, 2024, 53(3): 399-412. DOI: . doi: 10.11947/j.AGCS.2024.20230562
[3]	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.
[4]	戴激光, 王杨, 杜阳, 等. 光学遥感影像道路提取的方法综述[J]. 遥感学报, 2020, 24(7): 804-823.
	Dai Jiguang, Wang Yang, Du Yang, et al. Development and prospect of road extraction method for optical remote sensing image[J]. Journal of Remote Sensing, 2020, 24(7): 804-823.
[5]	Lian Renbao, Wang Weixing, Mustafa N, et al. Road extraction methods in high-resolution remote sensing images: a comprehensive review[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2020, 13: 5489-5507.
[6]	寇曼曼, 王勤忠, 万娜. 栅格地图道路的快速矢量化方法研究与实现[J]. 计算机与现代化, 2014(9): 82-85.
	Kou Manman, Wang Qinzhong, Wan Na. Study and implementation of fast vectorization for raster map road[J]. Computer and Modernization, 2014(9): 82-85.
[7]	刘帝佑, 孔赟珑, 陈静波, 等. 制图级矢量要素遥感智能解译的研究进展与展望[J]. 地球信息科学学报, 2025, 27(2): 285-304.
	Liu Diyou, Kong Yunlong, Chen Jingbo, et al. Recent progress and prospects in remote sensing image intelligent extraction of cartographic-level vector elements[J]. Journal of Geo-Information Science, 2025, 27(2): 285-304.
[8]	Chen Xin, Sun Qun, Guo Wenyue, et al. GA-Net: a geometry prior assisted neural network for road extraction[J]. International Journal of Applied Earth Observation and Geoinformation, 2022, 114: 103004.
[9]	Yu Lequan, Zhang Zhicheng, Li Xiaomeng, et al. Deep sinogram completion with image prior for metal artifact reduction in CT images[J]. IEEE Transactions on Medical Imaging, 2021, 40(1): 228-238.
[10]	Chen Xin, Yu Anzhu, Sun Qun, et al. Updating road maps at city scale with remote sensed images and existing vector maps[J]. IEEE Transactions on Geoscience and Remote Sensing, 2024, 62: 5616521.
[11]	陈奕州, 韦春桃. 遥感影像道路断线修复方法研究[J]. 地理空间信息, 2024, 22(4): 34-38.
	Chen Yizhou, Wei Chuntao. Repair method of road broke in remote sensing image[J]. Geospatial Information, 2024, 22(4): 34-38.
[12]	Zhang Xinchang, Shi Qian, Sun Ying, et al. The review of land use/land cover mapping AI methodology and application in the era of remote sensing big data[J]. Journal of Geodesy and Geoinformation Science, 2024, 7(3): 1-23.
[13]	郭威, 王立辉, 李洪堃. 基于圆形模板的遥感影像道路半自动提取方法[J]. 测绘与空间地理信息, 2021, 44(S1): 183-186, 191.
	Guo Wei, Wang Lihui, Li Hongkun. Semi-automatic road extraction method from remote sensing image based on circular template[J]. Geomatics & Spatial Information Technology, 2021, 44(S1): 183-186, 191.
[14]	连仁包, 王卫星, 李娟. 自适应圆形模板及显著图的高分辨遥感图像道路提取[J]. 测绘学报, 2018, 47(7): 950-958. DOI: . doi: 10.11947/j.AGCS.2018.20170596
	Lian Renbao, Wang Weixing, Li Juan. Road extraction from high-resolution remote sensing images based on adaptive circular template and saliency map[J]. Acta Geodaetica et Cartographica Sinica, 2018, 47(7): 950-958. DOI: . doi: 10.11947/j.AGCS.2018.20170596
[15]	Miao Zelang, Wang Bin, Shi Wenzhong, et al. A semi-automatic method for road centerline extraction from VHR images[J]. IEEE Geoscience and Remote Sensing Letters, 2014, 11(11): 1856-1860.
[16]	连仁包, 张振敏, 廖一鹏, 等. 结合测地距离场与曲线平滑的遥感图像道路中心线快速提取[J]. 测绘学报, 2023, 52(8): 1317-1329. DOI: . doi: 10.11947/j.AGCS.2023.20220002
	Lian Renbao, Zhang Zhenmin, Liao Yipeng, et al. A quick road centreline extraction method from remote sensing images combining with geodesic distance field and curve smoothing[J]. Acta Geodaetica et Cartographica Sinica, 2023, 52(8): 1317-1329. DOI: . doi: 10.11947/j.AGCS.2023.20220002
[17]	赵金奇, 杨杰, 李平湘, 等. 改进剖面匹配和EKF对SAR影像道路半自动提取[J]. 武汉大学学报(信息科学版), 2017, 42(8): 1144-1150.
	Zhao Jinqi, Yang Jie, Li Pingxiang, et al. Semi-automatic road extraction from SAR images using an improved profile matching and EKF[J]. Geomatics and Information Science of Wuhan University, 2017, 42(8): 1144-1150.
[18]	丁磊, 张保明, 郭海涛, 等. 矢量数据辅助的高分辨率遥感影像道路自动提取[J]. 遥感学报, 2017, 21(1): 84-95.
	Ding Lei, Zhang Baoming, Guo Haitao, et al. Automatic road extraction from high-resolution remote sensing images assisted by vector data[J]. National Remote Sensing Bulletin, 2017, 21(1): 84-95.
[19]	Vincze M, Kövári B. Comparative survey of thinning algorithms[C]//Proceedings of the 10th International Symposium of Hungarian Researchers on Computational Intelligence and Informatics. Budapest: Budapest Tech, 2009: 173-184.
[20]	Zhang T Y, Suen C Y. A fast parallel algorithm for thinning digital patterns[J]. Communications of the ACM, 1984, 27(3): 236-239.
[21]	陈品, 王涛, 张德港. 一种基于模板的二阶段ZS细化算法[J]. 计算机技术与发展, 2023, 33(2): 44-49, 56.
	Chen Pin, Wang Tao, Zhang Degang. A two-stage ZS thinning algorithm based on template[J]. Computer Technology and Development, 2023, 33(2): 44-49, 56.
[22]	韩建峰, 宋丽丽. 改进的字符图像细化算法[J]. 计算机辅助设计与图形学学报, 2013, 25(1): 62-66.
	Han Jianfeng, Song Lili. An improved thinning algorithm for character image[J]. Journal of Computer-Aided Design & Computer Graphics, 2013, 25(1): 62-66.
[23]	常庆贺, 吴敏华, 骆力明. 基于改进ZS细化算法的手写体汉字骨架提取[J]. 计算机应用与软件, 2020, 37(7): 107-113, 164.
	Chang Qinghe, Wu Minhua, Luo Liming. Handwritten Chinese character skeleton extraction based on improved ZS thinning algorithm[J]. Computer Applications and Software, 2020, 37(7): 107-113, 164.
[24]	Su Zhewen, Cao Zhongsheng, Wang Yuanzhen. Stroke extraction based on ambiguous zone detection: a preprocessing step to recover dynamic information from handwritten Chinese characters[J]. International Journal on Document Analysis and Recognition (IJDAR), 2009, 12(2): 109-121.
[25]	Cheng Tao, Zhang Yang, Haworth J. Network SpaceTime AI: concepts, methods and applications[J]. Journal of Geodesy and Geoinformation Science, 2022, 5(3): 78-92.
[26]	Fan D L, Wang B, Chen Z L, et al. Research on broken road connection method after road extraction from high-resolution remote sensing image[J]. The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 2020: 387-395.
[27]	Duan L Y, Mapurisa W, Leras M, et al. BrightEarth roads: towards fully automatic road network extraction from satellite imagery[C]//Proceedings of 2024 IEEE International Geoscience and Remote Sensing Symposium. Athens: IEEE, 2024: 9192-9196.
[28]	Teng Zixuan, Zheng Zezhong, Sun Xiangyang, et al. Topology-aware road extraction from remote sensing images using deep learning and graph-based connectivity refinement[J]. ISPRS International Journal of Geo-Information, 2026, 15(5): 208.
[29]	Deng Ligao, Deng Yupeng, Meng Yu, et al. GLD-Road: a global-local decoding road network extraction model for remote sensing images[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2025, 228: 741-755.
[30]	Bag S, Harit G. An improved contour-based thinning method for character images[J]. Pattern Recognition Letters, 2011, 32(14): 1836-1842.
[31]	周正扬, 詹恩奇, 郑建彬, 等. 局部关联度最优的手写汉字骨架提取[J]. 中国图象图形学报, 2017, 22(6): 833-841.
	Zhou Zhengyang, Zhan Enqi, Zheng Jianbin, et al. Skeleton extraction algorithm based on optimum local correlation degree for handwritten Chinese characters[J]. Journal of Image and Graphics, 2017, 22(6): 833-841.
[32]	Douglas D H, Peucker T K. Algorithms for the reduction of the number of points required to represent a digitized line or its caricature[J]. The Canadian Cartographer, 1973, 10(2): 112-122.
[33]	Demir I, Koperski K, Lindenbaum D, et al. DeepGlobe 2018: a challenge to parse the Earth through satellite images[C]//Proceedings of 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops. Salt Lake City: IEEE, 2018.
[34]	Mnih V. Machine learning for aerial image labeling[D]. Toronto: University of Toronto, 2013.
[35]	Chen L C, Zhu Yukun, Papandreou G, et al. Encoder-decoder with atrous separable convolution for semantic image segmentation[C]//Proceedings of 2018 Computer Vision. Cham: Springer International Publishing, 2018: 833-851.
[36]	Wu Honglin, Zhang Min, Huang Peng, et al. CMLFormer: CNN and multiscale local-context transformer network for remote sensing images semantic segmentation[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2024, 17: 7233-7241.
[37]	Zhou Lichen, Zhang Chuang, Wu Ming. D-LinkNet: LinkNet with pretrained encoder and dilated convolution for high resolution satellite imagery road extraction[C]//Proceedings of 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops. Salt Lake City: IEEE, 2018: 192-196.
[38]	Zhang Ke, Guo Wenyue, Yu Anzhu, et al. Road extraction network for remote sensing images based on multiscale feature fusion and edge-guided attention[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2025, 18: 24399-24414.

模型	方法	指标
模型	方法	S_N/m	S_R/m	N/（%）	T/s
CMLFormer	未修复	0.95	1.04	96.20	3.74
	断线修复法	0.93	1.01	96.52	4.51
	本文方法	0.89	0.99	96.89	4.39
DeeplabV3+	未修复	1.91	1.99	93.90	3.88
	断线修复法	1.90	1.97	94.52	4.12
	本文方法	1.82	1.91	94.92	3.86
DLinkNet34	未修复	1.10	1.18	96.21	3.70
	断线修复法	1.08	1.17	96.35	3.79
	本文方法	1.06	1.15	96.68	3.89
ME-Net	未修复	1.17	1.25	95.60	3.89
	断线修复法	1.12	1.21	95.92	3.77
	本文方法	1.08	1.15	96.56	3.71