Geographic location which can generally be described either absolutely or relatively is the locating information of a spatial object in geographic environment. Absolute method takes use of words or numbers that uniquely identifies the location of spatial entities, such as latitude and longitude coordinates, address, etc. by establishing a unified coding system. On the contrary, relative method locates a target object by the spatial relation between the target and its reference, which is coincident with people's spatial recognition habit and follows the development trend of socialization GIS service through combination with natural language. However, single sentence describing target location can cause such deficiencies as reliance on the very reference object and large uncertain locating area. This paper proposed a combinative description method to combine single description sentences of different reference objects and different spatial relations into one paragraph to approach the target location and diminish the reliance on the single reference object. Reference objects were cataloged to different types according to their characteristics: distinctive geographic semantics, a particular structure, and spatial relation constraints with the target object. Means of choosing various reference objects were discussed, and parameters of location information according to various reference objects such as spatial relations (including topological relation, direction relation and distance relation) as well as polylines' strike and polygons' extent were listed out, computed and transformed into corresponding words which were then filled into a sentence within the syntactic pattern. Under the principle of combination, single sentences were grouped into a comprehensive paragraph. The description experiment comparing single description sentence and combination paragraph showed that this approach attained various spatial recognition level and locating structure, and was preferable to better understand and locate a geographic location.
ZHENG Yue, LONG Yi, MING Xiaona, YAN Rui
. Natural Language Description of Geographic Location Considering Various Spatial Relations with Different References[J]. Journal of Geo-information Science, 2011
, 13(4)
: 465
-471
.
DOI: 10.3724/SP.J.1047.2011.00465
[1] Wang X, et al. A Study on How Humans Describe Relative Positions of Image Objects. Headway in Spatial Data Handling, Lecture Notes in Geoinformation and Cartography, 2008, 1-18.
[2] 李德仁, 李清泉, 谢智颖,等. 论空间信息与移动通信的集成应用[J]. 武汉大学学报(信息科学版), 2002,27(1):1-8.
[3] 马林兵, 龚健雅. 空间信息自然语言查询接口的研究与应用[J]. 武汉大学学报(信息科学版), 2003,28(3): 301-305.
[4] 余建伟, 李清泉. 位置感知计算中定位信息的自然语言描述[J]. 地理与地理信息科学, 2009, 25(1): 10-13, 21.
[5] Nothegger C, Winter S, Raubal M. Computation of the Salience of Features[M]. // Spatial Cognition and Computation, CiteSeer, 2004.
[6] 应申, 李霖, 高玉荣, 等. LBS中多尺度空间方位的自然语言描述[J]. 学术研究, 2008, 2(1):28-33.
[7] 龚咏喜, 刘瑜, 邬伦, 等. 基于带权 Voronoi图与地标的空间位置描述[J]. 地理与地理信息科学, 2010, 26(4): 21-26.
[8] Elias B. Extracting Landmarks with Data Mining Methods[M]. // Spatial Information Theory. Springer, 2003, 1-15.
[9] 张宏, 温永宁, 刘爱利, 等. 地理信息系统算法基础[M]. 北京:科学出版社, 2006.
[10] Egenhofer M, Franzosa R. Point-set Topological Spatial Relations[J]. International Journal of Geographical Information Systems, 1991,5(2): 161-174.
[11] 廖楚江, 杜清运. GIS空间关系描述模型研究综述[J]. 测绘科学, 2004, 29(4): 79-82,86.
[12] 杜世宏, 王桥, 李治江. GIS中自然语言空间关系定义[J]. 武汉大学学报(信息科学版), 2005, 30(6): 533-538.
[13] 刘新, 刘文宝, 李成名, 等. GIS中位置关系的定性描述及其推理[J]. 测绘科学技术学报, 2009, 26(2): 106-109.
[14] 杜冲, 司望利, 许珺. 基于地理语义的空间关系查询和推理[J]. 地球信息科学学报, 2010, 12(1): 48-55.
[15] 张雪英, 闾国年. 自然语言空间关系及其在 GIS中的应用研究[J]. 地球信息科学, 2007, 9(6): 77-81.
[16] 钱程扬, 龙毅, 徐震, 等. 基于 Web文本的开放式空间信息查询[J]. 武汉大学学报(信息科学版), 2010, 35(1): 83-87.
[17] 闫浩文, 郭仁忠. 空间方向关系基础性问题研究[J]. 测绘学报, 2002, 31(4):357-360.