结合SAM大模型和数学形态学的历史地图水系信息提取方法

doi:10.11947/j.AGCS.2024.20230308

测绘学报 ›› 2024, Vol. 53 ›› Issue (4): 761-772.doi: 10.11947/j.AGCS.2024.20230308

结合SAM大模型和数学形态学的历史地图水系信息提取方法

赵飞¹^,²(), 李兆正³, 甘泉⁴, 高祖瑜³, 王湛初³, 杜清运⁵, 王振声⁶, 沈洋⁴, 潘威⁷()

^1.云南大学地球科学学院，云南　昆明　650500
^2.云南省中老孟缅自然资源遥感监测国际联合实验室，云南　昆明　650051
^3.云南大学国际河流与生态安全研究院，云南　昆明　650500
^4.自然资源部第三大地测量队，四川　成都　610100
^5.武汉大学资源与环境科学学院，湖北　武汉　430079
^6.鹏城实验室战略与交叉前沿研究部，广东　深圳　518055
^7.云南大学历史与档案学院，云南　昆明　650091

收稿日期:2023-07-05 修回日期:2024-02-22 发布日期:2024-05-13
通讯作者: 潘威 E-mail:cartographer@ynu.edu.cn;panwei@ynu.edu.cn
作者简介:赵飞（1986—），男，博士，副教授，研究方向为地图学理论与应用、时空大数据分析。E-mail：cartographer@ynu.edu.cn
基金资助:
国家社科基金重大项目(22&ZD225);国家自然科学基金(41961064);四川省测绘地理信息局2023年新型基础测绘技术研究补助计划(2023KJ001);鹏城实验室重大项目(PCL2023AS6-1)

A method for hydrological information extraction from historical maps combining SAM large model and mathematical morphology

Fei ZHAO¹^,²(), Zhaozheng LI³, Quan GAN⁴, Zuyu GAO³, Zhanchu WANG³, Qingyun DU⁵, Zhensheng WANG⁶, Yang SHEN⁴, Wei PAN⁷()

^1.School of Earth Sciences, Yunnan University, Kunming 650500, China
^2.Yunnan International Joint Laboratory of China-Laos-Bangladesh-Myanmar Natural Resources Remote Sensing Monitoring, Kunming 650051, China
^3.Institute of International Rivers and Eco-Security, Yunnan University, Kunming 650500, China
^4.The Third Geodetic Surveying Brigade of Ministry of Natural Resources, Chengdu 610100, China
^5.School of Resource and Environmental Sciences, Wuhan University, Wuhan 430079, China
^6.Department of Strategic and Advanced Interdisciplinary Research, Peng Cheng Laboratory, Shenzhen 518055, China
^7.School of History and Archives, Yunnan University, Kunming 650091, China

Received:2023-07-05 Revised:2024-02-22 Published:2024-05-13
Contact: Wei PAN E-mail:cartographer@ynu.edu.cn;panwei@ynu.edu.cn
About author:ZHAO Fei (1986—), male, PhD, associate professor, majors in cartography theory & application and spatio-temporal big data analysis. E-mail: cartographer@ynu.edu.cn
Supported by:
The Major Program of the National Social Science Foundation of China(22&ZD225);The National Natural Science Foundation of China(41961064);The Sichuan Bureau of Surveying, Mapping and Geoinformation 2023 New Geodetic Technology Research Grant Program(2023KJ001);The Peng Cheng Laboratory Research Project(PCL2023AS6-1)

摘要/Abstract

摘要：

历史地图记载着丰富的历史地理信息，能够帮助了解历史规律，为当代发展提供借鉴。不同于现代地图、遥感影像等数据，历史地图保存时间久，存在留存数量少、图像精度低等问题，地图符号也与现代有所差异，因此信息难以被高效提取。针对该问题，本文以历史地图《宁夏省境黄河沿岸沟渠水道地形图》为试验数据，提出一种智能化历史地图水系信息提取方法。首先，结合符号句法，运用聚类与数学形态学方法构建数据集；然后，改进通用大模型（SAM）结构并进行迁移学习优化权重；最后，借助改进SAM自动提取历史地图水系信息。将试验结果与其他模型作对比，显示本文方法提取结果边界清晰，轮廓完整，准确率、精度等指标均为最高。同时，将提取结果与该区域水系现状作对比，发现历史上的河流沟渠如今大多改道、偏移或消失，湖泊面积大大减小。本文方法基于SAM通用大模型进行改进，验证了大模型在地图领域的可用性，为地图信息提取提供了思路。

关键词: 历史地图, 水系提取, 模糊C均值聚类, 数学形态学, SAM通用大模型

Abstract:

Historical maps record rich historical geographic information, which can help understand the laws of historical movement and provide reference for contemporary development. Different from modern maps, remote sensing images and other data, the historical map has been preserved for a long time, and there are some problems such as small number of reservations and low image accuracy. Map symbols are also different from modern maps, so the information is difficult to be extracted efficiently. Aiming at this problem, this study proposes an intelligent method for extracting hydrological information from historical maps based on the experimental data of topographic map of ditches and channels along the Yellow River in Ningxia province. Firstly, the datasets are constructed by clustering and mathematical morphology methods combined with symbolic syntax. Then, the general large model SAM structure is improved and the weight is optimized by transfer learning. Finally, the historical map hydrological information is automatically extracted by improved SAM. Comparing the experimental results with other models, it shows that the extraction results of this method have clear boundaries, complete contours, and the highest accuracy and accuracy. At the same time, the extraction results are compared with the current situation of the hydrological in the region. It is found that most of the rivers and ditches in history are now diverted, offset or disappeared, and the lake area is greatly reduced. The method in this paper is improved based on the SAM general large model, which verifies the availability of the large model in the map field and provides a new idea for map information extraction.

Key words: historical map, extraction of hydrology, fuzzy C-means, mathematical morphology, SAM general large model

中图分类号:

P283

赵飞, 李兆正, 甘泉, 高祖瑜, 王湛初, 杜清运, 王振声, 沈洋, 潘威. 结合SAM大模型和数学形态学的历史地图水系信息提取方法[J]. 测绘学报, 2024, 53(4): 761-772.

Fei ZHAO, Zhaozheng LI, Quan GAN, Zuyu GAO, Zhanchu WANG, Qingyun DU, Zhensheng WANG, Yang SHEN, Wei PAN. A method for hydrological information extraction from historical maps combining SAM large model and mathematical morphology[J]. Acta Geodaetica et Cartographica Sinica, 2024, 53(4): 761-772.

图/表 13

图1

图2

图3

图4

图5

图6

图7

图8

表1

图9

图10

表2

图11

参考文献 39

[1]	王光霞, 游雄, 於建峰, 等.地图设计与编绘[M].北京: 测绘出版社, 2014.
	WANG Guangxia, YOU Xiong, YU Jianfeng, et al. Map design and compilation[M]. Beijing: Surveying and Mapping Press, 2014.
[2]	翁敏, 黄谦, 苏世亮, 等. 基于皮尔斯符号三元观的专题地图符号设计[J]. 测绘地理信息, 2021, 46(1):44-47.
	WENG Min, HUANG Qian, SU Shiliang, et al. Design of thematic map symbols based on Peirce's semiotic triangle[J]. Journal of Geomatics, 2021, 46(1):44-47.
[3]	王家耀, 武芳, 闫浩文. 大变化时代的地图学[J]. 测绘学报, 2022, 51(6):829-842. DOI: 10.11947/j.AGCS.2022.20210661.
	WANG Jiayao, WU Fang, YAN Haowen. Cartography: its past, present and future[J]. Acta Geodaetica et Cartographica Sinica, 2022, 51(6):829-842. DOI: 10.11947/j.AGCS.2022.20210661.
[4]	赵飞, 杜清运, 任福, 等. 专题地图符号的句法结构及其自动构建机制[J]. 测绘学报, 2014, 43(6):653-660. DOI: 10.13485/j.cnki.11-2089.2014.0105.
	ZHAO Fei, DU Qingyun, REN Fu, et al. Syntactic characteristics and smart construction mechanism of thematic map symbol[J]. Acta Geodaetica et Cartographica Sinica, 2014, 43(6):653-660. DOI: 10.13485/j.cnki.11-2089.2014.0105.
[5]	PAL N R, PAL S K. A review on image segmentation techniques[J]. Pattern Recognition, 1993, 26(9):1277-1294.
[6]	ZHONG D X, YAN H. Color image segmentation using color space analysis and fuzzy clustering[C]//Proceedings of 2000 IEEE Signal Processing Society Workshop. Sydney: IEEE, 2000: 624-633.
[7]	LIU Tiange, XU Pengfei, ZHANG Shihui. A review of recent advances in scanned topographic map processing[J]. Neurocomputing, 2019, 328:75-87.
[8]	LEVACHKINE S, VELÀZQUEZ A, ALEXANDROV V, et al. Semantic analysis and recognition of raster-scanned color cartographic images[C]//Proceedings of 2001 International Workshop on Graphics Recognition. Berlin: Springer, 2001: 178-189.
[9]	KHOTANZAD A, BOUARFA A. Image segmentation by a parallel, non-parametric histogram based clustering algorithm[J]. Pattern Recognition, 1990, 23(9):961-973.
[10]	于子娟. 基于模糊C均值聚类的彩色图像分割方法研究[D]. 长沙: 中南大学, 2009.
	YU Zijuan. Research on color image segmentation method based on fuzzy C-means clustering[D]. Changsha: Central South University, 2009.
[11]	郑华利, 周献中, 王建宇. 基于色彩空间转换及模糊约束聚类的地形图自动分色[J]. 测绘学报, 2003, 32(2):183-187.
	ZHENG Huali, ZHOU Xianzhong, WANG Jianyu. Automatic color segmentation of topographic maps based on color space transformation and fuzzy restraint clustering[J]. Acta Geodaetica et Cartographic Sinica, 2003, 32(2):183-187.
[12]	LIM Y W, LEE S U. On the color image segmentation algorithm based on the thresholding and the fuzzy C-means techniques[J]. Pattern Recognition, 1990, 23(9):935-952.
[13]	林开颜, 徐立鸿, 吴军辉. 快速模糊C均值聚类彩色图像分割方法[J]. 中国图象图形学报, 2004, 9(2):159-163.
	LIN Kaiyan, XU Lihong, WU Junhui. A fast fuzzy C-means clustering for color image segmentation[J]. Journal of Image and Graphics, 2004, 9(2):159-163.
[14]	郑华利, 周献中, 王建宇. 空间关系信息和颜色信息相结合的地形图分层算法[J]. 中国图象图形学报, 2003, 8(3):334-340.
	ZHENG Huali, ZHOU Xianzhong, WANG Jianyu. Automatic color segmentation of topographic maps based on the combination of spatial relation information and color information[J]. Journal of Image and Graphics, 2003, 8(3):334-340.
[15]	KHOTANZAD A, ZINK E. Contour line and geographic feature extraction from USGS color topographical paper maps[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2003, 25(1):18-31.
[16]	GHIRCOIAS T, BRAD R. Contour lines extraction and reconstruction from topographic maps[J]. Ubiquitous Computing and Communication Journal, 2011, 6(2):681-691.
[17]	ZHANG Xiaoyu, CHEN Jiusheng, DONG Jiankang. Color clustering using self-organizing maps[C]//Proceedings of 2007 International Conference on Wavelet Analysis and Pattern Recognition. Beijing: IEEE, 2007: 986-989.
[18]	WANG Ming, YU Zhiwei, GUO Chengman, et al. Study on the method of generating digital elevation topographic map in the MATLAB[J]. Advanced Materials Research, 2012, 594/595/596/597:2902-2905.
[19]	MIAO Qiguang, XU Pengfei, LIU Tiange, et al. Linear feature separation from topographic maps using energy density and the shear transform[J]. IEEE Transactions on Image Processing, 2013, 22(4):1548-1558.
[20]	田宽. 彩色地形图中的点状符号识别算法研究[D]. 西安: 西安电子科技大学, 2017.
	TIAN Kuan. Research on point symbols recognition in color topographic map[D]. Xi'an: Xidian University, 2017.
[21]	崔腾腾, 刘纪平, 罗安. 一种卷积神经网络的地图图片智能识别方法[J]. 测绘科学, 2019, 44(1):118-123.
	CUI Tengteng, LIU Jiping, LUO An. Intelligent identification method of network map images based on convolutional neural network[J]. Science of Surveying and Mapping, 2019, 44(1):118-123.
[22]	任加新, 刘万增, 李志林, 等. 利用卷积神经网络进行“问题地图” 智能检测[J]. 武汉大学学报（信息科学版）, 2021, 46(4):570-577.
	REN Jiaxin, LIU Wanzeng, LI Zhilin, et al. Intelligent detection of “problematic map” using convolutional neural network[J]. Geomatics and Information Science of Wuhan University, 2021, 46(4):570-577.
[23]	李想. 基于深度学习的地图点注记识别及配置研究[D]. 郑州: 郑州大学, 2022.
	LI Xiang. Research on map point annotation recognition and configuration based on deep learning[D].Zhengzhou: Zhengzhou University, 2022.
[24]	周翰林. 基于深度学习的GIS图例符号识别系统研建[D]. 北京: 北京林业大学, 2018.
	ZHOU Hanlin. Study on GIS similar legend recognition system based on deep learning[D]. Beijing: Beijing Forestry University, 2018.
[25]	闫晓婧, 闫浩文, 王小龙, 等. 微地图符号的智能化匹配：“识图配符”[J]. 测绘科学, 2022, 47(10):205-212.
	YAN Xiaojing, YAN Haowen, WANG Xiaolong, et al. An intelligent symbol matching solution for we-maps through image recognition[J]. Science of Surveying and Mapping, 2022, 47(10):205-212.
[26]	ZHANG C, LIU L, CUI Y, et al. A comprehensive survey on segment anything model for vision and beyond[EB/OL]. [2023-06-10].https://arxiv.org/pdf/2305.08196.pdf.
[27]	张良培, 张乐飞, 袁强强. 遥感大模型：进展与前瞻[J]. 武汉大学学报（信息科学版）, 2023, 48(10):1574-1581.
	ZHANG Liangpei, ZHANG Lefei, YUAN Qiangqiang. Large remote sensing model: progress and prospects[J]. Geomatics and Information Science of Wuhan University, 2023, 48(10):1574-1581.
[28]	WANG Yi, ALBRECHT C M, ZHU Xiaoxiang. Self-supervised vision transformers for joint SAR-optical representation learning[EB/OL]. [2023-06-10]. https://arxiv.org/pdf/2204.05381.pdf.
[29]	REED C J, GUPTA R, LI Shufan, et al. Scale-MAE: a scale-aware masked autoencoder for multiscale geospatial representation learning[C]///Proceedings of 2023 IEEE/CVF International Conference on Computer Vision. Paris: IEEE, 2023: 4088-4099.
[30]	KIRILLOV A, MINTUN E, RAVI N, et al. Segment anything[EB/OL]. [2023-06-10].https://doi.org/10.48550/arXiv.2304.02643.
[31]	JIE L, ZHANG H. When SAM meets shadow detection[EB/OL]. [2023-06-10].https://arxiv.org/pdf/2305.11513.pdf.
[32]	PAL N R, BEZDEK J C. On cluster validity for the fuzzy c-means model[J]. IEEE Transactions on Fuzzy Systems, 1995, 3(3):370-379.
[33]	ZHAO Fei, DU Qingyun, REN Fu, et al. Syntactic characteristics and a smart construction mechanism for thematic map symbols[J]. Journal of Geodesy and Geoinformation Science, 2019, 2(1):37-48.
[34]	RIPLEY B D, MATHERON G. Random sets and integral geometry[J]. Journal of the Royal Statistical Society, 1976, 139(2):277-278.
[35]	SERRA J. Introduction to mathematical morphology[J]. Computer Vision, Graphics, and Image Processing, 1986, 35(3):283-305.
[36]	DOSOVITSKIY A, BEYER L, KOLESNIKOV A, et al. An image is worth 16×16 words: transformers for image recognition at scale[EB/OL]. [2023-06-10]. https://arxiv.org/pdf/2010.11929.pdf.
[37]	VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems. Long Beach: ACM Press, 2017: 6000-6010.
[38]	LIN T Y, GOYAL P, GIRSHICK R, et al. Focal loss for dense object detection[C]//Proceedings of 2017 IEEE International Conference on Computer Vision. Venice: IEEE, 2017: 2980-2988.
[39]	ZHAO Yao, WANG Guangxia, YANG Jian, et al. Building block extraction from historical maps using deep object attention networks[J]. ISPRS International Journal of Geo-Information, 2022, 11(11):572.

标签类别	数量	占比/（%）
沟渠	186	88.57
河流	59	28.10
浅湖	151	71.90
深湖	33	15.71
地图印章	15	7.14
文字注记	210	100.00
等高线	210	100.00
道路	91	43.30
居民地	118	56.19

模型	准确率	IoU	精度	F₁值
U-Net	84.31	61.62	62.32	76.48
原始SAM	89.68	76.65	84.70	84.40
改进SAM	94.13	83.25	91.18	90.86

结合SAM大模型和数学形态学的历史地图水系信息提取方法

A method for hydrological information extraction from historical maps combining SAM large model and mathematical morphology

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

图/表 13

参考文献 39

相关文章 5

编辑推荐

Metrics

本文评价

[1]	徐君, 王彩玲, 王丽. 采用PPI算法改进的一种数学形态学端元提取方法[J]. 测绘学报, 2019, 48(8): 996-1003.
[2]	林祥国, 张继贤. 面向对象的形态学建筑物指数及其高分辨率遥感影像建筑物提取应用[J]. 测绘学报, 2017, 46(6): 724-733.
[3]	刘祥磊, 童小华, 马静. 视频测量影像序列椭圆形人工目标点快速识别和跟踪方法[J]. 测绘学报, 2015, 44(6): 663-669.
[4]	隋立春,杨耘. 基于Car(p,q)模型和数学形态学理论的LiDAR点云数据滤波[J]. 测绘学报, 2012, 41(2): 219-224.
[5]	隋立春. 基于改进的数学形态学算法的LiDAR点云数据滤波[J]. 测绘学报, 2010, 39(4): 0-440.