Voronoi邻近关系支持下的点模式趋同提取方法

发布时间：2018-04-01 18:27

本文选题：点模式　切入点：Voronoi邻近关系　出处：《测绘学报》2017年05期

【摘要】：点模式及其趋同研究是揭示地学现象的产生、发展与演变,量化空间相似性分布、诠释空间分布成因的重要方式。目前,点模式研究侧重于已知频率与随机分布的一元独立性检验、距离测度下单观测值的二元相关性分析,而针对集聚过程相关性,空间拓扑与非拓扑邻近、综合多观测值的点模式趋同量化研究顾及不足。据此,以空间邻近性聚类、局部相关的多指标评价为切入点,本文提出了一种Voronoi邻近关系支持下的点模式趋同提取方法。首先,以Voronoi邻近相关表集聚算法剖分出空间独立性点模式;其次,依据Voronoi邻近关系指数测度、样本分布均值与分布方差的趋同假设,使用拉普拉斯平滑算子评价趋同度;最后,依据λ截矩阵,提取出Voronoi邻近、非Voronoi邻近关系支持下的强趋同点模式。试验以云南省腾冲市居民点数据为算例,经与点模式构建的聚类方法对比、趋同度计算与强趋同提取,验证了该方法的可行性与有效性。
[Abstract]:The study of point pattern and its convergence is an important way to reveal the origin, development and evolution of geoscience phenomena, quantify spatial similarity distribution and explain the causes of spatial distribution. The point model research focuses on the univariate independence test of known frequency and random distribution, the binary correlation analysis of the distance measure order observations, and the spatial topology and non-topological proximity to the correlation of the agglomeration process. The quantitative research of point pattern convergence based on multiple observations takes into account the deficiency. Based on spatial proximity clustering and local correlation multiple index evaluation, a method of point pattern convergence extraction supported by Voronoi proximity relationship is proposed in this paper. Secondly, according to the Voronoi neighbor relation index measure, the convergence hypothesis of sample distribution mean and distribution variance, Laplacian smoothing operator is used to evaluate the convergence degree. Based on 位 truncation matrix, the strong convergence point model supported by Voronoi proximity and non-#en1# proximity relationship is extracted. Taking the residential data of Tengchong City, Yunnan Province as an example, the convergence degree calculation and strong convergence extraction are compared with the clustering method constructed by the point model. The feasibility and effectiveness of the method are verified.
【作者单位】：中国矿业大学(北京)地球科学与测绘工程学院;昆明理工大学国土资源工程学院;国家基础地理信息中心;
【基金】：国家自然科学基金(41561082;41161061)~~
【分类号】：P208

【相似文献】