Douglas-Peucker算法全自动化的多尺度空间相似关系方法
王 荣(1986— ),女,甘肃天水人,博士生,主要从事多尺度空间相似关系研究。E-mail: freefly_99@126.com |
收稿日期: 2021-01-12
要求修回日期: 2021-03-12
网络出版日期: 2021-12-25
基金资助
国家自然科学基金项目(41930101)
国家自然科学基金青年基金项目(41801395)
甘肃省教育厅创新基金项目(2021A-110)
版权
Automation of the Douglas-Peucker Algorithm based on Spatial Similarity Relations
Received date: 2021-01-12
Request revised date: 2021-03-12
Online published: 2021-12-25
Supported by
Youth Foundation of National Natural Science Foundation of China(41930101)
Youth Foundation of National Natural Science Foundation of China(41801395)
Innovation Foundation of Education Department of Gansu(2021A-110)
Copyright
地图综合本质上是空间相似变换,研究Douglas-Peucker算法及其参数的设置,实质是研究算法的最佳距离阈值与尺度变化间的定量关系,但目前二者关系未知,导致参数设置及化简结果的选择主观性强。为此,提出以多尺度线要素空间相似关系为契合点,利用阈值参数寻优原理确定二者间定量关系,以实现基于DP算法线要素的全自动化简。结果表明:① 二次函数是描述最佳距离阈值与尺度变化间定量关系的最优函数;② 针对来源于相同地理特征区,如长江下游平原,的线要素可行,利用同一最佳距离阈值可实现基于DP算法线要素的全自动化简,且化简结果与已有成果数据吻合度较高;而来源于不同地理特征区域,如长江下游平原和江淮平原,的线要素,用同一最佳距离阈值化简是不合理。因此,应选择不同的最佳距离阈值,以实现不同地理特征区域线要素的DP算法全自动化简。
王荣 , 闫浩文 , 禄小敏 . Douglas-Peucker算法全自动化的多尺度空间相似关系方法[J]. 地球信息科学学报, 2021 , 23(10) : 1767 -1777 . DOI: 10.12082/dqxxkx.2021.210016
Map generalization is in essence a spatial similarity transformation of maps. Studying the Douglas-Peucker algorithm and its parameter setting is in essence studying the relationship between the optimal distance threshold of the algorithm and map scale change. However, the quantitative relationship between them is still unknown, which leads to strong subjectivity in parameter setting and selection of simplification results. Therefore, in order to realize the automated simplification of polyline based on DP algorithm, this paper proposes to take the spatial similarity evaluation model of multi-scale polylines as the coincidence point, and determine the quantitative relationship between them using the principle of threshold parameter optimization. The results indicate that quadratic function is the optimal function to describe the quantitative relationship between the optimal distance threshold and map scale change. It is feasible to use the same optimal distance to automatically simplify the polylines from the same geographical feature area based on the Douglas-Peucker algorithm, such as the polylines from the Lower Yangtze River plain. The simplification results match well with the existing target scale data. However, it is unreasonable to use the same optimal distance threshold to simplify the polylines from different geographical feature areas, such as polylines from the Lower Yangtze River plain and the Jianghuai plain. Therefore, different optimal distance thresholds should be selected to realize full automated simplification of DP algorithm for polylines from different geographical feature areas.
表1 多尺度线要素T寻优结果Tab. 1 Parameter T optimally results of multi-scale polylines |
山西山地 | T= -0.0001C2+0.0098C-0.037 (C (1,50]) | 长江 下游 平原 | T= -0.00003C2+0.0026C-0.0008 (C (1,50]) | |||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
C | 5 | 10 | 25 | 50 | C | 2.5 | 5 | 10 | 25 | 50 | ||
T/km | 0.0095 | 0.0510 | 0.1455 | 0.2030 | T/km | 0.0056 | 0.0114 | 0.0222 | 0.0455 | 0.0542 |
图12 化简结果几何特征及位置精度指标误差率Fig. 12 Error rate comparison of geometric and position accuracy of simplification results |
表2 化简结果几何精度评估Tab. 2 Geometric accuracy comparison of the simplification results |
几何指标 | 原比例尺数据(LM1) | 成果参照数据(LM2) | 本文的化简结果(L) | ||||||||
---|---|---|---|---|---|---|---|---|---|---|---|
边界线 | 道路 | 河流 | 边界线 | 道路 | 河流 | 边界线 | 道路 | 河流 | |||
长度/km | 198.9993 | 27.0565 | 153.7918 | 165.6861 | 26.1494 | 152.7020 | 165.6059 | 26.2379 | 148.6358 | ||
弯曲度 | 1.4493 | 1.5547 | 25.5883 | 1.1969 | 1.5103 | 25.0681 | 1.2067 | 1.5121 | 25.5663 |
表3 几何精度对比Tab. 3 Comparison of geometric and position accuracy |
线要素 | 成果参照数据 | 本文算法化简结果 | |||||||
---|---|---|---|---|---|---|---|---|---|
LR | SR | PE/km | SMD | LR′ | SR′ | PE′/km | SMD′ | ||
边界 | 0.8326 | 0.8258 | 0.2063 | 0.8070 | 0.8322 | 0.8324 | 0.1306 | 0.0998 | |
道路 | 0.9665 | 0.9714 | 0.0054 | 0.3323 | 0.9697 | 0.9725 | 0.0114 | 0.4565 | |
河流 | 0.9929 | 0.9796 | 0.0058 | 0.6046 | 0.9664 | 0.9991 | 0.0169 | 0.1203 |
[1] |
|
[2] |
|
[3] |
|
[4] |
|
[5] |
|
[6] |
|
[7] |
|
[8] |
钱海忠, 何海威, 王骁, 等. 采用三元弯曲组划分的线要素化简方法[J]. 武汉大学学报·信息科学版, 2017, 42(8):1096-1103.
[
|
[9] |
|
[10] |
|
[11] |
|
[12] |
|
[13] |
|
[14] |
王家耀, 何宗宜, 蒲英霞. 地图学[M]. 北京: 测绘出版社, 2016.
[
|
[15] |
|
[16] |
何海威, 钱海忠, 段佩祥, 等. 线要素化简及参数自动设置的案例推理方法[J]. 武汉大学学报·信息科学版, 2020, 45(3):344-352.
[
|
[17] |
|
[18] |
|
[19] |
|
[20] |
|
[21] |
|
[22] |
|
[23] |
武芳, 朱鲲鹏. 线要素化简算法几何精度评估[J]. 武汉大学学报·信息科学版, 2008, 33(6):600-603.
[
|
[24] |
顾腾, 陈晓勇, 刘成强. 一种Douglas-Peucker与Li-Openshaw结合改进的曲线化简方法[J]. 东华理工大学学报(自然科学版), 2016, 39(4):396-400.
[
|
[25] |
陈竞男, 钱海忠, 王骁, 等. 提高线要素匹配率的动态化简方法[J]. 测绘学报, 2016, 45(4):486-493.
[
|
[26] |
郭仁忠. 空间分析(2版)[M]. 北京: 高等教育出版社, 2001.
[
|
[27] |
|
[28] |
李成名, 郭沛沛, 殷勇, 等. 一种顾及空间关系约束的线化简算法[J]. 测绘学报, 2017, 46(4):498-506.
[
|
[29] |
|
/
〈 |
|
〉 |