数学基础是地图的重要要素,地图配准是恢复地图数学基础的主要方式,目前地图图像配准方法多关注于图 像特征的提取与匹配,未能有效利用图像特征点所对应的地理坐标信息。本文提出一种基于深度学习的特征点检测模 型与非线性变换模型结合的地图图像配准方法,训练并使用 YOLOv8-pose 模型高效提取图像中易于视觉识别且地理坐 标明确的特征点,利用其图像坐标和地理坐标代入加权最小二乘法求解非线性变换参数,实现无坐标地图图像的数学 基础恢复。结果表明,本文方法在面对多样化制图方式（如不同投影、比例尺、符号系统等）和复杂成像条件（如旋 转、透视畸变、噪声干扰等）影响下的地图图像,均能高效、精准地恢复其坐标系,配准精确率和召回率都超过 90%, 为地图图像融合其他含有坐标系的矢量与栅格数据,进行精确的空间知识理解、分析提供了新的技术思路。
[Objective] Mathematical foundations are integral to maps, enabling users to precisely interpret spatial relationships, feature positions, and geometric configurations. Restoring these foundations is critical for maps lacking coordinate information, particularly historical or digitized scans. Map registration serves as the primary means to align map images with standard geographic coordinates, facilitating their integration into geospatial analyses. However, existing registration methods prioritize image feature extraction over leveraging inherently associated geographic coordinates, compromising accuracy and robustness. This limitation is pronounced for maps with complex projections, variable scales, divergent symbology, or significant imaging distortions. Consequently, there is a pressing need for an adaptable, automated approach that harnesses semantic feature points with embedded geographic coordinates to restore cartographic mathematical frameworks.
[Method] This study introduces a map registration framework combining deep learning-based semantic keypoint detection with a cubic transformation model. A YOLOv8-pose architecture is trained on annotated data to efficiently identify visually discernible semantic keypoints while preserving their geographic coordinates. These paired image-geographic coordinates are then input into a weighted least squares algorithm to derive cubic transformation parameters, effectively modeling the spatial-to-geographic transformation. This process automates the recovery of mathematical foundations for unreferenced maps, minimizing manual intervention and enhancing resilience across diverse cartographic conditions.
[Result]Experiments validated the method's performance on six maps with varying projections, scales, symbology, and imaging artifacts (rotation, perspective distortion, overexposure, texture interference). The approach achieved over 90% precision and recall in semantic keypoint matching, demonstrating strong adaptability to challenging scenarios. By reconciling recovered mathematical frameworks with standard geographic data, the method successfully integrated unreferenced maps into hybrid geospatial datasets, vector, and raster formats.
[Conclusion] By integrating semantic features with geographic coordinates within a deep learning paradigm, this study achieves efficient, accurate, and robust restoration of map mathematical foundations. The proposed method addresses limitations of traditional approaches, such as rigid transformation models, heavy manual reliance, and poor generalizability. This work advances applications in historical cartography, thematic mapping, geospatial data fusion, and semantic geographic space analysis.
蓝秦隆, 邹进贵, 杨丁亮. 2022. 混合优化算法的点云配准. 测绘科学, 47(7): 119-125[Lan Q L, Zou J G, Yang D L.2022. Point cloud registration based on hybrid optimization algorithm. Science of Surveying and Mapping, 47(7): 119-125 (in Chinese)]
李国藻, 杨启和, 胡定荃. 1993. 地图投影. 北京:解放军出版社[Li G Z, Yang Q H, Hu D Q.1993.Map Projection. Beijing: People's Liberation Army Press (in Chinese)]
马瑞苑, 王勇, 车向红, 杜凯旋. 2024. 典型深度学习算法在地图图像特征点匹配上的适用性分析. 测绘科学, 49(8): 173-180[Ma R Y, Wang Y, Che X H, Du K X.2024. Analysis of the applicability of typical deep learning algorithms for map feature point matching. Science of Surveying and Mapping, 49(8): 173-180 (in Chinese)]
任福, 翁杰, 王昭, 张琛, 游泽威. 2022. 关于智能地图制图的几点思考. 武汉大学学报(信息科学版), 47(12):2064-2068[Ren F, Weng J, Wang Z, Zhang C, You Z W.2022. Some thoughts on smart cartography. Geomatics and Information Science of Wuhan University, 47(12):2064-2068(in Chinese)]
随银岭, 张宁, 秦志远, 童晓冲, 李贺, 赖广陵, 郭宇. 2020. 改进的基于累加投影图匹配的点云配准算法. 地理信息世界, 27(5): 17-22[Sui Y L, Zhang N, Qin Z Y, Tong X C, Li H, Lai G L, Guo Y.2020. An improved point cloud registration algorithm based on cumulative projection matching. Geomatics World, 27(5): 17-22 (in Chinese)]
王东峰. 2002. 多模态和大型图像配准技术研究. 博士学位论文. 北京:中国科学院电子学研究所[Wang D F. 2002. A study on registration of multi-modal and huge-size images. Doctoral Dissertation. Beijing: Aerospace Information Research Institute, Chinese Academy of Sciences(in Chinese)]
王家耀 .2017. 时空大数据时代的地图学 . 测绘学报 , 46(10): 1226-1237[Wang J Y.2017. Cartography in the age of spatio-temporal big data. Acta Geodaetica et Cartographica Sinica, 46(10): 1226-1237 (in Chinese)]
王铮, 刘纪平, 车向红, 王勇, 杜凯旋. 2022. 基于卷积神经网络的地图相似度匹配方法研究. 测绘科学, 47(7):169-175[Wang Z, Liu J P, Che X H, Wang Y, Du K X.2022. Research on map similarity matching method based on convolutional neural network. Science of Surveying and Mapping, 47(7):169-175(in Chinese)]
夏列钢, 杨存建, 陈军, 胡晓东. 2020. 目标网格编码支持的多源遥感影像地理匹配. 地理信息世界, 27(4): 36-40[Xia L G, Yang C J, Chen J, Hu X D.2020. A TGC-based approach for multi-source remote sensing image registration. Geomatics World, 27(4): 36-40 (in Chinese)]
叶沅鑫, 徐其志, 胡忠文. 2022. 多模态遥感图像配准. 北京: 科学出版社[Ye Y X, Xu Q Z, Hu Z W.2022. Multimodal Remote Sensing Image Registration. Beijing: Science Press (in Chinese)]
张泽瑞, 范大昭, 纪松, 董杨, 李东子, 刘杰. 2024. 结合注意力机制的卫星遥感影像立体匹配. 时空信息学报, 31(1): 41-49[Zhang Z R, Fan D Z, Ji S, Dong Y, Li D Z, Liu J.2024. Stereo matching of satellite remote sensing images based on attention mechanism. Journal of Spatio-temporal Information, 31(1): 41-49 (in Chinese)]
张智浩. 2022. 基于特征匹配的非刚性图像配准方法研究. 博士学位论文. 哈尔滨:哈尔滨工业大学[Zhang Z H.2022.Non-rigid image alignmment based on feature matching. Doctoral Dissertation. Harbin:Harbin Institute of Technology(in Chinese)]
周熙然, 李德仁, 薛勇, 汪云甲, 邵振峰. 2022. 地图图像智能识别与理解:特征、方法与展望. 武汉大学学报(信息科学版), 47(5): 641-650[Zhou X R, Li D R, Xue Y, Wang Y J, Shao Z F.2022. Intelligent map image recognition and understanding: Representative features, methodology and prospects, 47(5): 641-650 (in Chinese)]
邹展威, 赵东保, 张凯旋, 张克友. 2023. 面向国土空间规划“一张图”的国土资源专题地图图片的自动配准方法研究. 测绘工程, 32(2): 40-48[Zou Z W, Zhao D B, Zhang K X, Zhang K Y.2023.Research on the automatic registration method of land and resources thematic map pictures for “one map” of territorial spatial planning. Engineering of Surveying and Mapping, 32(2): 40-48 (in Chinese)]
DeTone D, Malisiewicz T, Rabinovich A.2017. SuperPoint: Self-supervised interest point detection and description. 1712.07629. https://arxiv.org/abs/1712.07629v4
Lowe D G.2004. Distinctive image features from scale-invariant keypoints. International Journal of Computer Vision, 60(2): 91-110
Maji D, Nagori S, Mathew M, Poddar D.2022. YOLO-pose: Enhancing YOLO for multi person pose estimation using object keypoint similarity loss//2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW). New Orleans, LA, USA. 2636-2645
Mo H L, Qu Y M, Li X D, Wang Z B.2024. Classification of liver tumors based on YOLOv8s-cls//2024 7th International Conference on Algorithms, Computing and Artificial Intelligence (ACAI). Guangzhou, China.1-5
Redmon J, Divvala S, Girshick R, Farhadi A.2016. You only look once: Unified, real-time object detection//2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Las Vegas, NV, USA. 779-788
Varghese R M S.2024. YOLOv8: A novel object detection algorithm with enhanced performance and robustness//2024 International Conference on Advances in Data Engineering and Intelligent Computing Systems (ADICS). Chennai, India. 1-6
Wang H Y, Li C Q, Li Z J, Du Y B, Zhou Z Q, Wu J Z.2024a. Breast ultrasound tumor detection based on improved YOLOv8s-OBB algorithm//2024 5th International Conference on Intelligent Computing and Human-Computer Interaction (ICHCI). Nanchang, China.120-125
Wang Y X, Huang Y J, Peng L Y, Wang M Y, Li W H, Jing M G, Zeng X Y.2024b. S-SIFT: A simple SIFT algorithm with high efficiency//2024 IEEE 17th International Conference on Solid-State & Integrated Circuit Technology (ICSICT), 1-3
Yang C J, Rong J.2024. Road segmentation algorithm based on improved YOLOv8-seg//2024 IEEE 4th International Conference on Electronic Technology, Communication and Information (ICETCI). Changchun, China. 1512-1516
Yang Z Q, Dan T T, Yang Y.2018. Multi-temporal remote sensing image registration using deep convolutional features. IEEE Access, 6: 38544-38555
