卷积神经网络支持下的建筑物选取方法

doi:10.11947/j.AGCS.2023.20220216

摘要/Abstract

摘要： 建筑物选取是地图综合的关键步骤,需要考虑目标大小、方向、形状、密度等多种上下文因子进行重要性评价与选取决策。现有方法大多考虑单一或少数几个因子,通过人工设置选取规则与参数,导致选取模型适应性不强。本文构建一种数据驱动的图卷积神经网络选取方法,该方法利用Delaunay三角网将建筑物目标组织为图结构,节点表示建筑物中心点,连接边体现建筑物之间的邻近关系,并计算建筑物的大小、方向、形状、密度特征作为对应节点的描述特征;然后叠加多个图傅里叶卷积运算构建图学习模型,并采用半监督学习方式训练模型,使之具备决策单个建筑物保留与否的能力。试验表明,本文方法能从少量的标注样本中有效地学习建筑物选取知识,在保留重要个体目标的同时也能较好地保持原有空间分布密度关系,克服传统方法在规则定义与参数设置方面的难题且不依赖于大量人工标注,为建筑物综合选取的智能化实施提供潜在的技术途径。

关键词: 地图综合, 建筑物选取, 图卷积神经网络, 半监督学习

Abstract: As a fundamental aspect of map generalization, building selection requires thorough consideration of various contextual factors such as size, orientation, shape, density, and more. However, many existing methods have only focused on single or a few factors, often relying on manual selection rules and parameters, which limits their practicality. In this study, we propose a data-driven building selection method using the graph convolutional network(GCN). Our method organizes buildings into a graph using Delaunay triangulation, with nodes representing building centers and edges denoting adjacent relationships between buildings. The size, orientation, shape, and density of each building are computed as the descriptive features for the associated nodes. Further, a GCN is constructed by stacking multiple graph Fourier convolutions and trained with semi-supervised learning to enable it to decide whether a building is selected or not. Experiments show that our method can effectively learn selection knowledge with few labels and perform well in maintaining the original spatial distribution densities and selecting important individual objects. It overcomes the difficulties in rule definition and parameter setting of traditional methods and does not rely on a large number of manual labels, which provides a promising solution for intelligent generalization.

Key words: map generalization, building selection, graph convolutional network, semi-supervised learning

中图分类号:

P283

安晓亚, 朱余德, 晏雄锋. 卷积神经网络支持下的建筑物选取方法[J]. 测绘学报, 2023, 52(9): 1574-1583.

AN Xiaoya, ZHU Yude, YAN Xiongfeng. A building selection method supported by graph convolutional network[J]. Acta Geodaetica et Cartographica Sinica, 2023, 52(9): 1574-1583.

参考文献

[1] 武芳, 巩现勇, 杜佳威. 地图制图综合回顾与前望[J]. 测绘学报, 2017, 46(10): 1645-1664. WU Fang, GONG Xianyong, DU Jiawei. Overview of the research progress in automated map generalization[J]. Acta Geodaetica et Cartographica Sinica, 2017, 46(10): 1645-1664.
[2] TÖPFER F, PILLEWIZER W. The principles of selection[J]. The Cartographic Journal, 1966, 3(1): 10-16.
[3] 杨敏. 顾及上下文特征的地图综合选取方法与应用研究[J]. 测绘学报, 2014, 43(8): 877. YANG Min. Research on feature selection considering spatial context in map generalization and itsapplication[J]. Acta Geodaetica et Cartographica Sinica, 2014, 43(8): 877.
[4] SADAHIRO Y. Cluster perception in the distribution of point objects[J]. Cartographica: the International Journal for Geographic Information and Geovisualization, 1997, 34(1): 49-62.
[5] REGNAULD N. Contextual building typification in automated map generalization[J]. Algorithmica, 2001, 30(2): 312-333.
[6] BURGHARDT D, CECCONI A. Mesh simplification for building typification[J]. International Journal of Geographical Information Science, 2007, 21(3): 283-298.
[7] 艾廷华, 刘耀林. 保持空间分布特征的群点化简方法[J]. 测绘学报, 2002, 31(2): 175-181. AI Tinghua, LIU Yaolin. A method of point cluster simplification with spatial distribution properties preserved[J]. Acta Geodaetica et Cartographic Sinica, 2002, 31(2): 175-181.
[8] 闫浩文, 王家耀. 基于Voronoi图的点群目标普适综合算法[J]. 中国图象图形学报, 2005, 10(5): 633-636. YAN Haowen, WANG Jiayao. A generic algorithm for point cluster generalization based on Voronoi diagrams[J]. Journal of Image and Graphics, 2005, 10(5): 633-636.
[9] 杨敏, 艾廷华, 卢威, 等. 自发地理信息兴趣点数据在线综合与多尺度可视化方法[J]. 测绘学报, 2015, 44(2): 228-234. DOI: 10.11947/j.AGCS.2015.20130564. YANG Min, AI Tinghua, LU Wei, et al. A real-time generalization and multi-scale visualization method for POI data in volunteered geographic information[J]. Acta Geodaetica et Cartographica Sinica, 2015, 44(2): 228-234. DOI: 10.11947/j.AGCS.2015.20130564.
[10] GONG Xianyong, WU Fang. A typification method for linear pattern in urban building generalisation[J]. Geocarto International, 2018, 33(2): 189-207.
[11] WANG X, BURGHARDT D. A typification method for linear building groups based on stroke simplification[J]. Geocarto International, 2021, 36(15): 1732-1751.
[12] WANG X, BURGHARDT D. A mesh-based typification method for building groups with grid patterns[J]. ISPRS International Journal of Geo-Information, 2019, 8(4): 168.
[13] 蔡永香, 郭庆胜. 基于Kohonen网络的点群综合研究[J]. 武汉大学学报(信息科学版), 2007, 32(7): 626-629. CAI Yongxiang, GUO Qingsheng.Points group generalization based on Konhonen net[J]. Geomatics and Information Science of Wuhan University, 2007, 32(7): 626-629.
[14] SESTER M. Optimization approaches for generalization and data abstraction[J]. International Journal of Geographical Information Science, 2005, 19(8/9): 871-897.
[15] 邓红艳, 武芳, 钱海忠, 等. 基于遗传算法的点群目标选取模型[J]. 中国图象图形学报, 2003, 8(8): 970-976. DENG Hongyan, WU Fang, QIAN Haizhong, et al. A model of point cluster selection based on genetical gorithms[J]. Journal of Image and Graphics, 2003, 8(8): 970-976.
[16] WANG Lin, GUO Qingsheng, LIU Yuangang, et al. Contextual building selection based on a genetic algorithm in map generalization[J]. ISPRS International Journal of Geo-Information, 2017, 6(9): 271.
[17] YAN Xiongfeng, CHEN Huan, HUANG Haoran, et al. Building typification in map generalization using affinity propagation clustering[J]. ISPRS International Journal of Geo-Information, 2021, 10(11): 732.
[18] 吕峥, 孙群, 马京振, 等. 复杂网络视角下的居民地选取方法[J]. 测绘学报, 2023, 52(5): 852-862. DOI: 10.11947/j.AGCS.2023.20220267. LV Zheng, SUN Qun, MA Jingzhen, et al. Residential area selection method from the perspective of complex network[J]. Acta Geodaetica et Cartographica Sinica, 2023, 52(5): 852-862. DOI: 10.11947/j.AGCS.2023.20220267.
[19] SHEN Yilang, AI Tinghua, ZHAO Rong. Raster-based method for building selection in the multi-scale representation of two-dimensional maps[J]. Geocarto International, 2022, 37(22): 6494-6518.
[20] 程涛, 张洋, James Haworth. 基于网络和图的时空智能: 概念、方法和应用[J]. 测绘学报, 2022, 51(7): 1629-1639. DOI: 10.11947/j.AGCS.2022.20220236. CHENG Tao, ZHANG Yang, HAWORTH J. Network and graph-based SpaceTimeAI: conception, method and applications[J]. Acta Geodaetica et Cartographica Sinica, 2022, 51(7): 1629-1639. DOI: 10.11947/j.AGCS.2022.20220236.
[21] YANG Min, JIANG Chenjun, YAN Xiongfeng, et al. Detecting interchanges in road networks using a graph convolutional network approach[J]. International Journal of Geographical Information Science, 2022, 36(6): 1119-1139.
[22] YAN Xiongfeng, AI Tinghua, YANG Min, et al. A graph convolutional neural network for classification of building patterns using spatial vector data[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2019, 150: 259-273.
[23] YANG Min, KONG Bo, DANG Ruirong, et al. Classifying urban functional regions by integrating buildings and points-of-interest using a stacking ensemble method[J]. International Journal of Applied Earth Observation and Geoinformation, 2022, 108: 102753.
[24] YAN Xiongfeng, AI Tinghua, YANG Min, et al. Graph convolutional autoencoder model for the shape coding and cognition of buildings in maps[J]. International Journal of Geographical Information Science, 2021, 35(3): 490-512.
[25] 于洋洋, 贺康杰, 武芳, 等. 面状居民地形状分类的图卷积神经网络方法[J]. 测绘学报, 2022, 51(11): 2390-2402. DOI: 10.11947/j.AGCS.2022.20210134. YU Yangyang, HE Kangjie, WU Fang, et al. Graph convolution neural network method for shape classification of areal settlements[J]. Acta Geodaetica et Cartographica Sinica, 2022, 51(11): 2390-2402. DOI: 10.11947/j.AGCS.2022.20210134.
[26] 晏雄锋, 袁拓, 杨敏, 等. 建筑物形状特征分析表达与自适应化简方法[J]. 测绘学报, 2022, 51(2): 269-278. DOI: 10.11947/j.AGCS.2022.20210317. YAN Xiongfeng, YUAN Tuo, YANG Min, et al. An adaptive building simplification approach based on shape analysis and representation[J]. Acta Geodaetica et Cartographica Sinica, 2022, 51(2): 269-278. DOI: 10.11947/j.AGCS.2022.20210317.
[27] DUCHÊNE C, BARD S, BARILLOT X, et al. Quantitative and qualitative description of building orientation [C]//Proceedings of the 7th ICA Workshop on Progress in Automated Map Generalisation. Paris: International Cartographic Association Commission on Map Generalization, 2003.
[28] BASARANER M, CETINKAYA S. Performance of shape indices and classification schemes for characterising perceptual shape complexity of building footprints in GIS[J]. International Journal of Geographical Information Science, 2017, 31(10): 1952-1977.
[29] KIPF T N, WELLING M. Semi-supervised classification with graph convolutional networks[C]//Proceedings of 2017 International Conference on Learning Representations. Toulon: [s.n.], 2017.
[30] YANG Min, YUAN Tuo, YAN Xiongfeng, et al. A hybrid approach to building simplification with an evaluator from a backpropagation neural network[J]. International Journal of Geographical Information Science, 2022, 36(2): 280-309.