-
摘要:
机器学习在滑坡的易发性评价中面临两个难点，一是评价指标的客观量化，二是训练样本的选择。鉴于此，采用频率比法实现了评价指标的客观量化，利用k均值聚类算法实现了非滑坡样本数据的筛选。结果表明，以k均值聚类算法筛选非滑坡为前提，神经网络的训练精度由73%提升到了97%，支持向量机的训练精度由75%提升到了96%。基于GIS平台，将神经网络和支持向量机模型计算的全区易发性指数按自然断点法分为五个区域，分区图与历史灾害点的叠加分析统计结果显示，神经网络在全局范围内的评价结果优于支持向量机模型，全局精度分别为76%和74%。研究结果可为南江县的防灾减灾工作提供参考。
Abstract:Machine learning faces two difficulties in the evaluation of landslide susceptibility. One is the objective quantification of evaluation index, and the other is the selection of training sample-0.5pts. For that reason, the frequency ratio method is used to achieve the objective quantification of evaluation index, and the k-means clustering algorithm is used to achieve the selection of non-landslide sample data. The results show that based on the premise that the k-means clustering algorithm selects non-landslides, the training accuracy of the neural network has increased from 73% to 97%, and the training accuracy of the support vector machine has increased from 75% to 96%. Based on the GIS platform, the susceptibility index calculated by the neural network and support vector machine model is divided into five regions according to the natural break point method. The statistical results of the overlay analysis of the zoning map and the historical disaster points show that the evaluation result of the neural network is better than the support vector machine model in the global scope, and the global accuracy is 76% and 74%, respectively. The research results can provide reference for disaster prevention and mitigation in Nanjiang County of China.
-
-
表 1 数据源
Table 1. Data source
数据名称 数据类型 数据来源 滑坡灾害点 excel 南江县1∶5万地质灾害详查 DEM 栅格 地理空间数据云 1∶25万地质图 栅格 91卫图 
下载: 导出CSV
表 2 因子量化结果
Table 2. Results of factor quantification
因子 因子二级属性 si ni xi 坡度/(°) 0～10 650642 67 1.085425029 10～16 803628 107 1.403446432 16～22 805749 96 1.255852346 22～29 763630 53 0.731576946 29～34 375968 20 0.560720245 34～43 311624 15 0.507373154 >43 72850 1 0.144689684 坡向 北 438393 28 0.673227029 东北 416896 37 0.935494243 东 450786 51 1.192523317 东南 479436 49 1.077289835 南 510916 50 1.031543684 西南 491470 60 1.286830544 西 499440 39 0.823092053 西北 496754 45 0.954856841 坡型 <−1 131994 1 0.079856989 −1 ～ −0.3 832505 66 0.835649597 −0.3 ～ 0.1 1141012 127 1.173223173 0. 1 ～ 0.8 1491162 153 1.081517936 >0.8 187418 12 0.674896336 水系/m <200 331210 40 1.272986136 200～500 479405 64 1.407163424 500～800 455006 48 1.111965305 800～1200 558891 65 1.225895254 1200～1500 381262 34 0.939988453 1500～2000 548717 51 0.979690471 2000～5000 1023859 57 0.586815838 >5000 5741 0 0.000000000 岩组 K1 882610 101 1.206200914 J3 1024962 120 1.234072302 J1-2 266052 77 3.050642528 T1-2 223391 22 1.038064004 P2-3 94493 2 0.223098927 Pz1 360578 21 0.613885241 Z 426090 14 0.346332954 ξγNh 326208 2 0.064625291 Pt2 148769 0 0.000000000 S1-2 30938 0 0.000000000 高程/m 332 ～ 604 424239 84 2.087064249 604 ～ 768 713900 121 1.786549738 768 ～ 924 647443 77 1.253592279 924 ～ 1094 538137 44 0.861840594 1094 ～1273 451639 23 0.536788894 1273 ～ 1464 377209 7 0.195606426 1464 ～ 1670 276415 3 0.114400197 1670 ～ 1897 236877 0 0.000000000 1897 ～ 2493 118232 0 0.000000000 地形起伏/m 9～98 543341 64 1.241579747 98～151 997907 120 1.267530155 151～202 929250 105 1.191033159 202～258 720598 53 0.775264576 258～329 447958 13 0.305895564 329～655 145037 4 0.290702192
下载: 导出CSV
表 3 因子相关性分析结果
Table 3. Results of factor correlation analysis
因子 坡度 坡向 坡型 水系 岩组 高程 地形 坡度 1.00 0.02 0.15 −0.05 0.16 0.16 0.61 坡向 0.02 1.00 0.02 0.01 0.01 0.00 0.00 坡型 0.15 0.02 1.00 0.01 0.09 0.09 0.17 水系 −0.05 0.01 0.01 1.00 0.12 0.13 −0.06 岩组 0.16 0.01 0.09 0.12 1.00 0.19 0.25 高程 0.16 0.00 0.09 0.13 0.19 1.00 0.27 地形 0.61 0.00 0.17 −0.06 0.25 0.27 1.00
下载: 导出CSV
表 4 k均值聚类统计分析结果
Table 4. Results of k-means clustering statistical analysis
聚类结果 栅格数量 滑坡点数 相对滑坡比 0 909306 37 0.428902710 1 836114 106 1.336310845 2 963716 28 0.306249991 3 265935 77 3.051984680 4 809020 111 1.446208281
下载: 导出CSV
表 5 神经网络分区统计结果
Table 5. Partition statistics results of neural network
易发性等级 栅格数量 分区面积比例/% 滑坡点数 相对滑坡频率比 不易发 1333655 35.24 35 0.276625155 低易发 654221 17.29 38 0.612246399 中易发 738454 19.51 70 0.999175361 高易发 626759 16.56 103 1.732222873 极高易发 431002 11.39 113 2.763543349
下载: 导出CSV
表 6 支持向量机分区统计结果
Table 6. Partition statistical results of support vector machines
易发性等级 栅格数量 分区面积比例/% 滑坡点数 相对滑坡频率比 不易发 1125524 29.74 43 0.402699248 低易发 943452 24.93 49 0.547448656 中易发 830829 21.96 72 0.913456715 高易发 463728 12.25 87 1.977529889 极高易发 420558 11.11 108 2.706854924
下载: 导出CSV
-
[1] 邱海军. 区域滑坡崩塌地质灾害特征分析及其易发性和危险性评价研究: 以宁强县为例[D]. 西安: 西北大学, 2012.
QIU Haijun. Study on the regional landslide characteristic analysis and hazard assessment: A case study of Ningqiang County[D]. Xi'an: Northwest University, 2012. (in Chinese with English abstract)
[2] 唐亚明, 张茂省, 李林, 等. 滑坡易发性危险性风险评价例析[J]. 水文地质工程地质,2011,38(2):125 − 129. [TANG Yaming, ZHANG Maosheng, LI Lin, et al. Discrimination to the landslide susceptibility, hazard and risk assessment[J]. Hydrogeology & Engineering Geology,2011,38(2):125 − 129. (in Chinese with English abstract) doi: 10.3969/j.issn.1000-3665.2011.02.022
[3] 杨秀梅. 基于GIS的地质灾害危险性评价[D]. 兰州: 兰州大学, 2008.
YANG Xiumei. Geological hazard risk assessment based on GIS[D]. Lanzhou: Lanzhou University, 2008. (in Chinese with English abstract)
[4] 胡芹龙, 王运生. 基于GIS的川西地貌过渡带滑坡灾害易发性评价[J]. 成都理工大学学报(自然科学版),2018,45(6):746 − 753. [HU Qinlong, WANG Yunsheng. The susceptibility assessment of geological disasters in geomorphic transition zone based on GIS, western Sichuan, China[J]. Journal of Chengdu University of Technology (Science & Technology Edition),2018,45(6):746 − 753. (in Chinese with English abstract)
[5] 郭长宝, 唐杰, 吴瑞安, 等. 基于证据权模型的川藏铁路加查: 朗县段滑坡易发性评价[J]. 山地学报,2019,37(2):240 − 251. [GUO Changbao, TANG Jie, WU Ruian, et al. Landslide susceptibility assessment based on WOE model along Jiacha—Langxian County section of Sichuan—Tibet railway, China[J]. Mountain Research,2019,37(2):240 − 251. (in Chinese with English abstract)
[6] 兰恒星, 伍法权, 王思敬. 基于GIS的滑坡CF多元回归模型及其应用[J]. 山地学报,2002,20(6):732 − 737. [LAN Hengxing, WU Faquan, WANG Sijing. GIS based landslide CF multi-variable regression model and its application[J]. Journal of Mountain Research,2002,20(6):732 − 737. (in Chinese with English abstract) doi: 10.3969/j.issn.1008-2786.2002.06.015
[7] 郭子正, 殷坤龙, 黄发明, 等. 基于滑坡分类和加权频率比模型的滑坡易发性评价[J]. 岩石力学与工程学报,2019,38(2):287 − 300. [GUO Zizheng, YIN Kunlong, HUANG Faming, et al. Evaluation of landslide susceptibility based on landslide classification and weighted frequency ratio model[J]. Chinese Journal of Rock Mechanics and Engineering,2019,38(2):287 − 300. (in Chinese with English abstract)
[8] 许冲, 戴福初, 姚鑫, 等. GIS支持下基于层次分析法的汶川地震区滑坡易发性评价[J]. 岩石力学与工程学报,2009,28(增刊 2):3978 − 3985. [XU Chong, DAI Fuchu, YAO Xin, et al. Gis-based landslide susceptibility assessment using analytical hierarchy process in Wenchuan earthquake region[J]. Chinese Journal of Rock Mechanics and Engineering,2009,28(Sup 2):3978 − 3985. (in Chinese with English abstract)
[9] 苏强. 基于DEM的黄土滑坡危险性评价研究[D]. 北京: 中国地质大学(北京), 2006.
SU Qiang. Research on loess landslide hazard zonation based on DEM[D]. Beijing: China University of Geosciences, 2006. (in Chinese with English abstract)
[10] 李嘉良, 马东辉, 王威. 基于证据理论和熵权灰色关联的潜在地震滑坡危险性评价[J]. 中南大学学报(自然科学版),2016,47(5):1730 − 1736. [LI Jialiang, MA Donghui, WANG Wei. Assessment of potential seismic landslide hazard based on evidence theory and entropy weight grey incidence[J]. Journal of Central South University (Science and Technology),2016,47(5):1730 − 1736. (in Chinese with English abstract) doi: 10.11817/j.issn.1672-7207.2016.05.036
[11] 牛瑞卿, 彭令, 叶润青, 等. 基于粗糙集的支持向量机滑坡易发性评价[J]. 吉林大学学报(地球科学版),2012,42(2):430 − 439. [NIU Ruiqing, PENG Ling, YE Runqing, et al. Landslide susceptibility assessment based on rough sets and support vector machine[J]. Journal of Jilin University (Earth Science Edition),2012,42(2):430 − 439. (in Chinese with English abstract)
[12] 王森, 许强, 罗博宇, 等. 基于分形理论的南江县滑坡敏感性分析与易发性评价[J]. 水文地质工程地质,2017,44(3):119 − 126. [WANG Sen, XU Qiang, LUO Boyu, et al. Vulnerability analysis and susceptibility evaluation of landslides based on fractal theory in Nanjiang County[J]. Hydrogeology & Engineering Geology,2017,44(3):119 − 126. (in Chinese with English abstract)
[13] 何静. 基于机器学习的滑坡灾害空间预测及风险评估[D]. 成都: 电子科技大学, 2019.
HE Jing. Spatial prediction and risk assessment of landslides based on machine learning[D]. Chengdu: University of Electronic Science and Technology of China, 2019. (in Chinese with English abstract)
[14] 黄发明, 殷坤龙, 蒋水华, 等. 基于聚类分析和支持向量机的滑坡易发性评价[J]. 岩石力学与工程学报,2018,37(1):156 − 167. [HUANG Faming, YIN Kunlong, JIANG Shuihua, et al. Landslide susceptibility assessment based on clustering analysis and support vector machine[J]. Chinese Journal of Rock Mechanics and Engineering,2018,37(1):156 − 167. (in Chinese with English abstract)
[15] 王敞, 陈增强, 袁著祉. 基于遗传算法的K均值聚类分析[J]. 计算机科学,2003,30(2):163 − 164. [WANG Chang, CHEN Zengqiang, YUAN Zhuzhi. K-means clustering based on genetic algorithm[J]. Computer Science,2003,30(2):163 − 164. (in Chinese with English abstract) doi: 10.3969/j.issn.1002-137X.2003.02.044
[16] 陈玉萍, 袁志强, 周博, 等. 遗传算法优化BP网络在滑坡灾害预测中的应用研究[J]. 水文地质工程地质,2012,39(1):114 − 119. [CHEN Yuping, YUAN Zhiqiang, ZHOU Bo, et al. Application of back propagation neural networks with optimization of genetic algorithms to landslide hazard prediction[J]. Hydrogeology & Engineering Geology,2012,39(1):114 − 119. (in Chinese with English abstract)
[17] 武雪玲, 沈少青, 牛瑞卿. GIS支持下应用PSO-SVM模型预测滑坡易发性[J]. 武汉大学学报(信息科学版),2016,41(5):665 − 671. [WU Xueling, SHEN Shaoqing, NIU Ruiqing. Landslide susceptibility prediction using GIS and PSO-SVM[J]. Geomatics and Information Science of Wuhan University,2016,41(5):665 − 671. (in Chinese with English abstract)
[18] 张俊, 殷坤龙, 王佳佳, 等. 三峡库区万州区滑坡灾害易发性评价研究[J]. 岩石力学与工程学报,2016,35(2):284 − 296. [ZHANG Jun, YIN Kunlong, WANG Jiajia, et al. Evaluation of landslide susceptibility for Wanzhou district of Three Gorges Reservoir[J]. Chinese Journal of Rock Mechanics and Engineering,2016,35(2):284 − 296. (in Chinese with English abstract)
-
图(6)
表(6)
计量
- 文章访问数: 2696
- PDF下载数: 109
- 施引文献: 0