A reinforcement learning method for collaborative generalization of soundings and depth contours

doi:10.11947/j.AGCS.2024.20230084

Abstract

Abstract:

Nowadays, the existing methods of automatic cartographic generalization usually generalize soundings and depth contours separately, which easily leads to unsatisfactory generalization results. To address this problem, a reinforcement learning method for collaborative generalization of soundings and depth contours is proposed. Firstly, training samples for collaborative generalization are obtained. Simultaneously, a reinforcement learning model is constructed based on the cartographic constraints and the related algorithms. Then, the constructed model is trained by using the sample data, so that the interaction between soundings and depth contours can be explored in the generalization process. Finally, the generalization algorithms of soundings and depth contours can be adaptively adjusted by utilizing the trained model, so that the mutual influence relationship between soundings and depth contours can be fully considered in the generalization process. The experimental results show that: compared with current common methods, the proposed method can effectively improve the quality of the cartographic generalization results, and is more suitable for the collaborative generalization of soundings and depth contours.

Key words: nautical cartography, submarine topographical generalization, automatic sounding selection, automatic depth contour simplification, reinforcement learning

CLC Number:

P208

Zikang SONG, Shuaidong JIA, Zhicheng LIANG, Lihua ZHANG, Chuan LIANG. A reinforcement learning method for collaborative generalization of soundings and depth contours[J]. Acta Geodaetica et Cartographica Sinica, 2024, 53(7): 1345-1354.

Figures/Tables 12

Fig.1

Fig.2

Fig.3

Fig.4

Tab.1

Tab.2

Tab.3

Tab.4

References 28

[1]	张立华, 贾帅东, 唐露露, 等. 航海图自动制图综合理论与方法[M]. 北京: 科学出版社, 2022: 25, 85.
	ZHANG Lihua, JIA Shuaidong, TANG Lulu, et al. Theory and methods of automatic cartographic generalizing for nautical chart [M]. Beijing: Science Press, 2022: 25, 85.
[2]	张立华, 李改肖, 郑义东, 等. 海图制图综合[M]. 北京: 国防工业出版社, 2017: 118, 130.
	ZHANG Lihua, LI Gaixiao, ZHENG Yidong, et al. Cartographic generalization [M]. Beijing: National Defense Industry Press, 2017: 118, 130.
[3]	国家质量技术监督局. 中国航海图编绘规范：GB 12320—1998[S]. 北京: 中国标准出版社, 1999.
	State Bureau of Quality and Technical Supervision of the People's Republic of China. Specifications for Chinese nautical charts: GB 12320—1998[S]. Beijing: Standards Press of China, 1999.
[4]	武芳, 杜佳威, 吴芳华. 海底地貌数据综合研究进展[J]. 测绘学报, 2022, 51(7):1588-1605. DOI: 10.11947/j.AGCS.2022.20220119.
	WU Fang, DU Jiawei, WU Fanghua. A survey of the research progress in automated generalization of submarine terrain data[J]. Acta Geodaetica et Cartographica Sinica, 2022, 51(7):1588-1605. DOI: 10.11947/j.AGCS.2022.20220119.
[5]	YAN Haowen, WEIBEL R. An algorithm for point cluster generalization based on the Voronoi diagram[J]. Computers & Geosciences, 2008, 34(8):939-954.
[6]	文佳昕, 李靖涵, 行瑞星, 等. 顾及地形复杂度的多波束测深数据抽稀算法[J]. 测绘科学技术学报, 2018, 35(4):435-440.
	WEN Jiaxin, LI Jinghan, XING Ruixing, et al. A thinning algorithm of multibeam sounding data considering the complexity of seabed terrain[J]. Journal of Geomatics Science and Technology, 2018, 35(4):435-440.
[7]	贾帅东, 张立华, 董箭, 等. 利用拟构模型面调控水深模型点选取的航海DDM构建[J]. 武汉大学学报(信息科学版), 2019, 44(11):1715-1722.
	JIA Shuaidong, ZHANG Lihua, DONG Jian, et al. A method for constructing DDM serving for navigation using pre-constructed model surface to control and adjust the selection of DDM nodes[J]. Geomatics and Information Science of Wuhan University, 2019, 44(11):1715-1722.
[8]	贾帅东, 张立华, 彭认灿, 等. 确保水深模型航海安全性的深度保证率控制方法[J]. 交通运输工程学报, 2015, 15(5):101-109.
	JIA Shuaidong, ZHANG Lihua, PENG Rencan, et al. Control method of probability of adequate depth ensuring navigation safety of depth model[J]. Journal of Traffic and Transportation Engineering, 2015, 15(5):101-109.
[9]	张立华, 贾帅东, 王涛, 等. 一种不漏水深浅点的海图水深自动选取方法: CN103955932B[P]. 2016-08-24.
	ZHANG Lihua, JIA Shuaidong, WANG Tao, et al. Automatic charted depth selection method of water-leakage free depth points: CN103955932B[P]. 2016-08-24.
[10]	贾帅东. 航海DDM的构建理论与方法[D]. 大连: 海军大连舰艇学院, 2015.
	JIA Shuaidong. The theories and methods for constructing DDM serving for navigation [D]. Dalian: Dalian Naval Academy, 2015.
[11]	SKOPELITI A, STAMOU L, TSOULOS L, et al. Generalization of soundings across scales: from DTM to harbour and approach nautical charts[J]. ISPRS International Journal of Geo-Information, 2020, 9(11):693.
[12]	YU Linyi, DU Jiawei, ZHAI Renjian, et al. A fast generalization method of multibeam echo soundings for nautical charting[J]. Journal of Geovisualization and Spatial Analysis, 2021, 6(1):2.
[13]	高王军, 彭认灿, 陈轶, 等. 双向缓冲区算法及其在海图线状要素综合中的应用研究[J]. 测绘科学, 2009, 34(2):187-190.
	GAO Wangjun, PENG Rencan, CHEN Yi, et al. The research of the double direction buffering algorithm and its application on the generalizing of chart linear feature[J]. Science of Surveying and Mapping, 2009, 34(2):187-190.
[14]	李靖涵, 武芳, 杜佳威, 等. Delaunay三角网支持下的海图等深线化简[J]. 武汉大学学报(信息科学版), 2019, 44(5):778-783.
	LI Jinghan, WU Fang, DU Jiawei, et al. Chart depth contour simplification based on Delaunay triangulation[J]. Geomatics and Information Science of Wuhan University, 2019, 44(5):778-783.
[15]	李靖涵. 海底地貌自动制图综合算法研究[D]. 郑州: 信息工程大学, 2018.
	LI Jinghan. Research on automated cartographic generalization of submarine relief [D]. Zhengzhou: Information Engineering University, 2018.
[16]	SKOPELITI A, TSOULOS L, PE'ERI S. Depth contours and coastline generalization for harbour and approach nautical charts[J]. ISPRS International Journal of Geo-Information, 2021, 10(4):197.
[17]	魏泓丞, 张立华, 唐露露, 等. 顾及局部差异的航海图线要素复杂度度量[J]. 测绘学报, 2022, 51(9):1959-1968. DOI: 10.11947/j.AGCS.2022.20210129.
	WEI Hongcheng, ZHANG Lihua, TANG Lulu, et al. Complexity measurement of line element on nautical chart considering local difference[J]. Acta Geodaetica et Cartographica Sinica, 2022, 51(9):1959-1968. DOI: 10.11947/j.AGCS.2022.20210129.
[18]	贾帅东, 张立华, 彭认灿. 与等深线协调的水深自动选取方法[J]. 哈尔滨工程大学学报, 2016, 37(1):59-64.
	JIA Shuaidong, ZHANG Lihua, PENG Rencan. Method for automatic selection of water depth harmonizing with a depth contour[J]. Journal of Harbin Engineering University, 2016, 37(1):59-64.
[19]	YAN Jingya, GUILBERT E, SAUX E. An ontology-driven multi-agent system for nautical chart generalization[J]. Cartography and Geographic Information Science, 2017, 44(3):201-215.
[20]	YAN Jingya, GUILBERT E, SAUX E. An ontology of the submarine relief for analysis and representation on nautical charts[J]. The Cartographic Journal, 2015, 52(1):58-66.
[21]	何海威, 钱海忠, 王骁, 等. 采用弯曲进行道路化简冲突避免的方法[J]. 测绘学报, 2016, 45(3):354-361. DOI: 10.11947/j.AGCS.2016.20140576.
	HE Haiwei, QIAN Haizhong, WANG Xiao, et al. Avoiding special conflicts in road simplification by using road bends[J]. Acta Geodaetica et Cartographica Sinica, 2016, 45(3):354-361. DOI: 10.11947/j.AGCS.2016.20140576.
[22]	WANG Yanjun, LI Shaochun, WANG Mengjie, et al. A simple deep learning network for classification of 3D mobile LiDAR point clouds[J]. Journal of Geodesy and Geoinformation Science, 2021, 4(3):49-59.
[23]	何海威, 钱海忠, 谢丽敏, 等. 立交桥识别的CNN卷积神经网络法[J]. 测绘学报, 2018, 47(3):385-395. DOI: 10.11947/j.AGCS.2018.20170265.
	HE Haiwei, QIAN Haizhong, XIE Limin, et al. Interchange recognition method based on CNN[J]. Acta Geodaetica et Cartographica Sinica, 2018, 47(3):385-395. DOI: 10.11947/j.AGCS.2018.20170265.
[24]	KRIZHEVSKY A, SUTSKEVER I, HINTON G E. ImageNet classification with deep convolutional neural networks[J]. Communications of the ACM, 2017, 60(6):84-90.
[25]	杜佳威, 武芳, 朱丽, 等. 图形、图像融合利用的集成学习智能化简方法及其在岛屿岸线化简中的应用[J]. 测绘学报, 2022, 51(3):373-387. DOI: 10.11947/j.AGCS.2022.20210135.
	DU Jiawei, WU Fang, ZHU Li, et al. An ensemble learning simplification approach based on multiple machine-learning algorithms with the fusion using of raster and vector data and a use case of coastline simplification[J]. Acta Geodaetica et Cartographica Sinica, 2022, 51(3):373-387. DOI: 10.11947/j.AGCS.2022.20210135.
[26]	GOODFELLOW I, BENGIO Y, COURVILLE A. Deep learning [M]. Cambridge: MIT Press, 2016: 72.
[27]	张立华, 温连发, 贾帅东. 航海图等深线化简质量的定量评估指标与方法[J]. 武汉大学学报(信息科学版), 2018, 43(4):496-501, 508.
	ZHANG Lihua, WEN Lianfa, JIA Shuaidong. Indices and methods for evaluating quantificationally the quality of simplification of a depth-contour in nautical chart[J]. Geomatics and Information Science of Wuhan University, 2018, 43(4):496-501, 508.
[28]	张立华, 贾帅东, 王涛, 等. 深度保证率和表达度指标的定义及评估方法[J]. 武汉大学学报(信息科学版), 2015, 40(5):695-700.
	ZHANG Lihua, JIA Shuaidong, WANG Tao, et al. Definitions and estimating methods of a probability of an adequate depth and representiveness[J]. Geomatics and Information Science of Wuhan University, 2015, 40(5):695-700.

数据集	训练集	测试集
平均奖励	0.873	0.836
均方根误差	0.008	0.047

综合方法	安全性指标		图理性指标	准确性指标		美观性指标	运行效率指标
综合方法	扩深长度/cm	是否漏浅	信息表达是否存在矛盾	深度表达度/（%）	偏浅量/m	协调度/（%）	时间消耗/s
简单综合法	0	否	是	85.79	1.02	66.8	4.33
冲突避免法	1.67	否	否	85.42	1.04	72.5	4.35
协调水深法	0	否	否	77.34	2.12	86.6	6.56
强化学习法	0	否	否	83.14	1.21	78.4	5.06

综合方法	安全性指标		图理性指标	准确性指标		美观性指标	运行效率指标
综合方法	扩深长度/cm	是否漏浅	信息表达是否存在矛盾	深度表达度/（%）	偏浅量/m	协调度/（%）	时间消耗/s
简单综合法	0	否	是	91.92	0.31	57.3	7.02
冲突避免法	2.24	否	否	91.89	0.31	62.4	7.04
协调水深法	0	否	否	87.63	0.55	84.0	11.19
强化学习法	0	否	否	91.61	0.32	86.7	9.69

综合方法	安全性指标		图理性指标	准确性指标		美观性指标	运行效率指标
综合方法	扩深长度/cm	是否漏浅	信息表达是否存在矛盾	深度表达度/（%）	偏浅量/m	协调度/（%）	时间消耗/s
简单综合法	0	否	是	87.80	2.13	58.1	6.61
冲突避免法	0.56	否	否	87.12	2.25	58.2	6.63
协调水深法	0	否	否	80.29	3.45	80.7	10.44
强化学习法	0	否	否	85.74	2.49	84.3	8.50