基于层次匹配的面状要素几何匹配方法研究
作者简介:张旗升(1990-),硕士生,研究方向为GIS方法和应用。E-mail:275534955@qq.com
收稿日期: 2015-10-13
要求修回日期: 2015-11-30
网络出版日期: 2016-08-10
基金资助
国家自然科学基金项目(41371375)
北京市自然科学基金项目(8132018)
The Research of Polygon Geometric Matching Method Based on Hierarchical Matching
Received date: 2015-10-13
Request revised date: 2015-11-30
Online published: 2016-08-10
Copyright
空间数据增量信息提取是空间数据库增量更新的基础,而要素几何匹配则为空间数据增量信息提取的关键流程之一。针对面状要素几何匹配时所产生的单一匹配算子匹配准确度低,以及多算子加权匹配算法匹配效率低、冗余度高、结果不准确等情况,本文提出一种基于多层次匹配的面状要素几何匹配算法。该算法以反映面状要素位置、大小、形状特征的质心匹配算子、面积重叠度算子以及转向角函数匹配算子为基础,并按各算子复杂度进行层次匹配规则设计,以提高算法效率、保证匹配的准确性。本文通过2种方法进行对比实验,结果表明:层次匹配算法匹配完整度和准确度均能达到90%以上,可以很好地进行面状要素几何匹配;与加权匹配方法相比效率更高,代价更小,且所得匹配结果比加权匹配方法所得结果更加准确。
张旗升 , 王艳慧 . 基于层次匹配的面状要素几何匹配方法研究[J]. 地球信息科学学报, 2016 , 18(8) : 1043 -1051 . DOI: 10.3724/SP.J.1047.2016.01043
Incremental information extraction of geospatial data is the foundation of geospatial database incremental updating, and geometric matching of features is one of the key processes in the incremental information extraction of geospatial data. In order to solve the problems that the single matching operator couldn't perform accurately in the polygon geometric matching, and the negative impacts of low efficiency, redundant matching, low accuracy of weighted multi-operator matching algorithm, etc, a polygon geometric matching method based on hierarchical matching is proposed. To improve the efficiency of the algorithm and guarantee the accuracy of matching results, this algorithm was formulated based on the centroid matching operator, areal overlapping degree matching operator, and steering angle function matching operator which could reflect the location, size and shape feature of polygon respectively, and it was designed with the hierarchical matching rules according to the complexity of each operator. In this paper, a contrast experiment was carried out between two methods, the results show that the integrity and accuracy of the hierarchical matching method can be higher than 90%, and the method works better in the geometric matching of polygon. Compared with the weighted matching method, this method has a high efficiency and low cost; at the same time, the result is more accurate than the weighted matching method.
Tab.1 The change types of planar entities and the analysis of corresponding geometrical features表1 面状实体变化类型以及几何特征分析 |
变化类型 | 几何特征分析 |
---|---|
属性变化 | 几何特征相同,属性不同 |
扩大 | 形状、位置相同,大小不同 |
缩小 | 形状、位置相同,大小不同 |
平移 | 形状、大小相同,位置不同 |
旋转 | 位置、大小相同,形状部分相同 |
变形 | 形状、大小、位置均不相同 |
Tab.2 Comparison of the experimental results表2 实验结果对比 |
匹配方法 | 要素样本数 | 匹配实验选取要素数/ | 匹配实验选取要素中实际匹配数/ | 匹配数据集中实际匹配数/ | 完整度/ (%) | 准确度/ (%) |
---|---|---|---|---|---|---|
层次匹配方法 | 312 | 251 | 243 | 264 | 92.0 | 96.8 |
加权匹配方法 | 312 | 256 | 231 | 264 | 87.5 | 90.0 |
Fig.1 Overall flow chart of the algorithm process图1 算法总体流程图示 |
Fig.2 Flow chart of hierarchical matching method图2 层次匹配具体流程图 |
Fig.3 Display of the experimental data图3 实验数据展示 |
Fig.4 A part of the hierarchical matching method′s results图4 层次匹配实验部分结果 |
Fig.5 A part of the weighted matching method′s results图5 加权匹配实验部分结果 |
Fig.6 Visualization of a part of the hierarchicalmatching method′s results图6 层次匹配法部分匹配结果 注:红色地物代表匹配地物;白色地物代表非匹配地物 |
Fig.7 Visualization of a part of the weighted matching method′s results注:红色地物代表匹配地物;白色地物代表非匹配地物 图7 加权匹配法部分匹配结果 |
Tab.3 Analysis of mistaken feature instances in the weighted matching method′s results表3 加权匹配方法错误要素实例分析 |
对应ID | 质心匹配 | 面积重叠度匹配 | 转向角匹配 | 加权算法 | 层次匹配结果 | 加权算法匹配结果 |
---|---|---|---|---|---|---|
MD1_32、MD2_30 | 0.999 | 0.979 | 0.513 | 0.830 | 否 | 是 |
MD1_33、MD2_32 | 0.998 | 0.965 | 0.495 | 0.819 | 否 | 是 |
MD1_65、MD2_47 | 0.781 | 0.753 | 0.986 | 0.840 | 否 | 是 |
Fig.8 Examples of mistaken feature instances in the weighted matching method's results图8 加权匹配方法错误要素实例展示 |
Tab.4 The Matching results of experiments in different study areas表4 不同研究区域匹配实验结果 |
研究区域序号 | 样本个数 | 层次匹配法完整度/(%) | 层次匹配法准确度/(%) | 加权匹配法完整度/(%) | 加权匹配法准确度/(%) |
---|---|---|---|---|---|
1 | 113 | 91.2 | 97.1 | 86.8 | 89.6 |
2 | 196 | 93.3 | 96.7 | 85.3 | 87.0 |
3 | 287 | 91.7 | 96.4 | 85.1 | 89.9 |
Tab.5 Comparison of operators′calculation time between the two methods表5 两种方法子算子计算次数对比 |
方法类型 | 质心匹配算子计算次数 | 面积重叠度匹配算子计算次数 | 转向角匹配算子计算次数 | ||
---|---|---|---|---|---|
加权匹配方法 | m×n | m×n | m×n | ||
层次匹配方法 | m×n | K1 | K2 |
The authors have declared that no competing interests exist.
[1] |
|
[2] |
|
[3] |
[
|
[4] |
[
|
[5] |
[
|
[6] |
[
|
[7] |
[
|
[8] |
[
|
[9] |
[
|
[10] |
[
|
[11] |
[
|
[12] |
|
[13] |
|
[14] |
[
|
[15] |
[
|
[16] |
[
|
[17] |
[
|
[18] |
|
[19] |
[
|
/
〈 | 〉 |