河南省滑坡灾害易发性制图研究：多种机器学习模型的对比

曹文庚; 潘登; 徐郅杰; 张文培; 任宇; 南天

doi:10.19509/j.cnki.dzkq.tb20230338

河南省滑坡灾害易发性制图研究：多种机器学习模型的对比

doi: 10.19509/j.cnki.dzkq.tb20230338

曹文庚^1, ,,
潘登^{2, 3},
徐郅杰^{2, 3},
张文培^{2, 3},
任宇¹,
南天¹

1.
中国地质科学院水文地质环境地质研究所，石家庄 050061
2.
河南省自然资源监测和国土整治院，郑州 450016
3.
河南省地质灾害防治重点实验室，郑州 450016

基金项目: 河南省重点研发与推广专项（科技攻关）（232102321012）；国家自然科学基金项目（41972262）；河北自然科学基金优秀青年科学基金项目（D2020504032）；河南省重点研发专项（221111321500）

详细信息

通讯作者:
E-mail：281084632@qq.com

中图分类号: P642.22
计量
- 文章访问数: 503
- PDF下载量: 85
- 被引次数: 0
出版历程
- 收稿日期: 2023-06-13
- 录用日期: 2023-12-29
- 修回日期: 2023-12-04
- 网络出版日期: 2023-12-29

Landslide disaster vulnerability mapping study in Henan Province: Comparison of different machine learning models

CAO Wengeng^{1
, ,},
PAN Deng^{2, 3},
XU Zhijie^{2, 3},
ZHANG Wenpei^{2, 3},
REN Yu¹,
NAN Tian¹

1.
Institute of Hydrogeology and Environmental Geology, CAGS, Shijiazhuang 050061, China
2.
Henan Provincial Institute of Natural Resources Monitoring and Land Consolidation with Rehabilitation, Zhengzhou 450016 , China
3.
Henan Provincial Key Laboratory of Geological Disaster Prevention and Control, Zhengzhou 450016 , China

More Information

Corresponding author: E-mail：281084632@qq.com

摘要

摘要:
河南省具有复杂的地貌类型，面临着频繁发生滑坡灾害的挑战，因此进行高效准确的滑坡易发性制图对于地方政府和居民具有重要意义。但是，在滑坡易发性制图研究中，如何选取适合河南省滑坡灾害数据集的机器学习模型、提高评价精度的对比研究仍需进一步开展。以河南省为研究区，收集滑坡数据并编录成滑坡灾害数据库。通过递归特征消除法筛选出对滑坡相对影响最高的11个因子（坡度、高程、平面曲率、剖面曲率、土地覆盖、岩性、土壤类型、降水量、道路密度、河流密度、断裂带密度）整合成空间数据集，训练多层感知机（MLP）神经网络、随机森林（RF）、极端梯度提升（XGBoost）和支持向量机（SVM）模型并使用接收者受试特征曲线下面积（AUC）评估各个模型性能，制作高精度滑坡易发性分区图。研究结果表明，多层感知机模型对河南省滑坡灾害数据集适配性最强，AUC达到0.95。相较于支持向量机、极端梯度提升和随机森林模型，MLP模型预测的滑坡灾害高易发区的面积占比最小，能更精确地识别潜在滑坡灾害高风险区域。预测的极高和高易发区主要分布在豫西山地、丘陵地区，地形因素对河南省滑坡灾害发育具有主导作用。研究成果可为大尺度区域开展高精度滑坡灾害易发性评价提供参考。
- 滑坡 /
- 多层感知机 /
- 机器学习 /
- 易发性制图 /
- 河南省
Abstract: Objective
Henan Province, with its complex geomorphology, and faces the challenge of frequent landslide disasters. Therefore, efficient and accurate landslide susceptibility mapping (LSM) is of great significance for local governments and residents. However, further comparative research is needed on how to select machine learning models suitable for the landslide disaster dataset in Henan Province to improve evaluation accuracy in landslide susceptibility mapping research.
Methods
This study takes Henan Province as the research area, collected landslide data and compiled it into a landslide disaster database. By using the recursive feature elimination method, the 11 factors that have the highest relative impact on landslides (slope, elevation, plan curvature, profile curvature, land cover, lithology, soil type, precipitation, road density, river density, fault density) were selected and integrated into a spatial dataset. Multi layer perceptron (MLP) neural network, random forest (RF), extreme gradient Boosting (XGBoost), and support vector machine (SVM) models were trained, and the model performances were evaluated with receiver operating characteristic curves and the area under the curve, finally, a high-precision landslide susceptibility zoning map was created.
Results
The MLP model showed the strongest adaptability to the landslide disaster dataset in Henan Province, with an AUC of 0.95. Compared to SVM, XGBoost, and RF models, the MLP model predicted the smallest landslide proportion in highly susceptible areas, and can more accurately identify potential high-risk areas for landslide disasters. The predicted extremely high and high-risk areas are mainly distributed in the mountainous and hilly areas of western Henan Province, and terrain factors play a dominant role in the development of landslide disasters in Henan Province.
Conclusion
These results provide a high-accuracy reference for landslide susceptibility assessment over large areas.
- landslide /
- multilayer perceptron /
- machine learning /
- susceptibility mapping /
- Henan Province

HTML全文

图 1 研究技术框架

Figure 1. Technical framework used this study

下载: 全尺寸图片幻灯片

图 2 研究区域内滑坡的位置

Figure 2. Locations of landslides in the study area

下载: 全尺寸图片幻灯片

图 3 滑坡影响因子

Figure 3. Landslide impact factors

下载: 全尺寸图片幻灯片

图 4 基于RFE的滑坡影响因子重要性排序

Figure 4. Importance ranking of landslide impact factors on the basis of RFE

下载: 全尺寸图片幻灯片

图 5 非滑坡点分布

Figure 5. Distribution map of nonlandslide points

下载: 全尺寸图片幻灯片

图 6 利用测试数据集分析不同滑坡模型的ROC曲线

Figure 6. Analysing the ROC curves of different landslide models by using test data sets

下载: 全尺寸图片幻灯片

图 7 基于SVM、XGBoost、RF和MLP模型的滑坡易发性概率图

Figure 7. Landslide susceptibility probability map based on the SVM, XGBoost, RF and MLP models

下载: 全尺寸图片幻灯片

图 8 基于SVM、XGBoost、RF和MLP模型的滑坡易发性图

Figure 8. Landslide susceptibility diagram based on the SVM, XGBoost, RF and MLP models

下载: 全尺寸图片幻灯片

表 1 滑坡点及影响因子的数据来源

Table 1. Data sources of landslide points and impact factors

滑坡影响因子	数据获取	数据来源
滑坡点	河南省地质灾害点分布数据	地理遥感生态网（http://www.gisrs.cn）
高程	河南省30 m精度DEM数字高程数据	地理空间数据云（http://www.gscloud.cn）
坡度
坡向
曲率
平面曲率
剖面曲率
土壤类型	河南省30 m精度土壤类型分布数据	HWSD土壤数据库
土地覆盖	河南省30 m精度土地覆盖数据	全球土壤覆盖数据库（http://www.globallandcover.com）
降水量	河南省30 a降水量数据	NCDC公开FTP服务器
岩性	河南省岩性数据	中国科学院资源环境科学数据中心（http://www.resdc.cn）
距道路距离	河南省30 m精度水系、交通、居民点和土地利用数据	全国地理信息资源目录服务系统（https://www.webmap.cn）
距河流距离
距断层距离
断裂带密度
道路密度
河流密度

下载: 导出CSV

表 2 滑坡易发性评估统计结果

Table 2. Statistical results of landslide susceptibility evaluation

预测模型	易发性等级面积占比/%
预测模型	极高	高	中	低	极低
SVM	15.63	9.90	5.76	7.08	61.63
XGBoost	23.98	3.57	2.77	3.90	65.78
RF	18.61	7.97	7.15	14.47	51.80
MLP	10.71	5.87	5.18	7.96	70.27

下载: 导出CSV

参考文献(35)

[1]	赖国泉,焦海平,吴红刚,等. 山区机场高填方滑坡特征、成灾机理、防治及启示:以攀枝花机场为例[J/OL]. 地质科技通报. [2024-12-04]. http://doi.org/10.19509/j.cnki.dzkq.tb20240216. LAI G Q,JIAO H P,WU H G,et al. Characteristics、disaster mechanism、prevention and treatment and enlighrenment of airport high fill landslide in mountainous area: Taking Panzhihua Airport as an example[J/OL]. Bulletin of Geological Science and Technology. [2024-12-04]. https://doi.org/10.19509/j.cnki.dzkq.tb20240216. (in Chinese with English abstract
[2]	BRABB E E. Innovative approaches to landslide hazard and risk mapping[C]//Anon. International landslide symposium proceedings. Toronto,Canada:[S. n.],1985:17-22.
[3]	于宪煜. 基于多源数据和多尺度分析的滑坡易发性评价方法研究[D]. 北京:中国地质大学(北京),2016. YU X Y. Research on landslide susceptibility evaluation method based on multi-source data and multi-scale analysis [D]. Beijing:China University of Geosciences(Beijing),2016. (in Chinese with English abstract
[4]	MORAGUES S,LENZANO M G,LANFRI M,et al. Analytic hierarchy process applied to landslide susceptibility mapping of the North Branch of Argentino Lake,Argentina[J]. Natural Hazards,2021,105(1):915-941. doi: 10.1007/s11069-020-04343-8
[5]	LI Y,WANG X,MAO H. Influence of human activity on landslide susceptibility development in the Three Gorges area[J]. Natural Hazards,2020,104:2115-2151. doi: 10.1007/s11069-020-04264-6
[6]	YOUSSEF A M,POURGHASEMI H R,POURTAGHI Z S,et al. Landslide susceptibility mapping using random forest,boosted regression tree,classification and regression tree,and general linear models and comparison of their performance at Wadi Tayyah Basin,Asir Region,Saudi Arabia[J]. Landslides,2016,13:839-856. doi: 10.1007/s10346-015-0614-1
[7]	田乃满,兰恒星,伍宇明,等. 人工神经网络和决策树模型在滑坡易发性分析中的性能对比[J]. 地球信息科学学报,2021,22(12):2304-2316. TIAN N M,LAN H X,WU Y M, et al. Comparison of the performance of artificial neural network and decision tree model in landslide susceptibility analysis[J]. Journal of Geo-Information Science,2021,22(12):2304-2316. (in Chinese with English abstract
[8]	HUANG F,CAO Z,JIANG S H,et al. Landslide susceptibility prediction based on a semi-supervised multiple-layer perceptron model[J]. Landslides,2020,17:2919-2930. doi: 10.1007/s10346-020-01473-9
[9]	PRAKASH N,MANCONI A,LOEW S. Mapping landslides on EO data:Performance of deep learning models vs. traditional machine learning models[J]. Remote Sens., 2020,12(3):346.
[10]	CHEN W,PENG J,HONG H,et al. Landslide susceptibility modelling using GIS-based machine learning techniques for Chongren County,Jiangxi Province,China[J]. Science of the Total Environment,2018,626:1121-1135. doi: 10.1016/j.scitotenv.2018.01.124
[11]	NGO P T T,PANAHI M,KHOSRAVI K,et al. Evaluation of deep learning algorithms for national scale landslide susceptibility mapping of Iran[J]. Geoscience Frontiers,2021,12(2):505-519. doi: 10.1016/j.gsf.2020.06.013
[12]	KAVZOGLU T,TEKE A. Predictive performances of ensemble machine learning algorithms in landslide susceptibility mapping using random forest,extreme gradient boosting (XGBoost) and natural gradient boosting (NGBoost)[J]. Arabian Journal for Science and Engineering,2022,47(6):7367-7385. doi: 10.1007/s13369-022-06560-8
[13]	李冬冬. 降雨诱发豫西锁固型滑坡演化机理模型试验研究[D]. 郑州:华北水利水电大学,2020. LI D D. Experimental study on the evolution mechanism of rainfall induced locked landslides in western Henan[D]. Zhengzhou:North China University of Water Resources and Electric Power,2020. (in Chinese with English abstract
[14]	VAPNIK V N. Statistical learning theory[M]. [S. 1.]:Wiley-Interscience,1998.
[15]	OOMMEN T,MISRA D,TWARAKAVI N K C,et al. An objective analysis of support vector machine based classification for remote sensing[J]. Mathematical Geosciences,2008,40(4):409-424. doi: 10.1007/s11004-008-9156-6
[16]	KAMRAN K V,FEIZIZADEH B,KHORRAMI B,et al. A comparative approach of support vector machine kernel functions for GIS-based landslide susceptibility mapping[J]. App. Geomat.,2021,13:837-851. doi: 10.1007/s12518-021-00393-0
[17]	PATLE A,CHOUHAN D S. SVM kernel functions for classification[C]//Anon. International Conference on Advances in Technology and Engineering (ICATE). [S. 1. ]:[S. n. ],2013:1-9.
[18]	ZHANG L,ZHANG B. Relationship between support vector set and kernel functions in SVM[J]. J. Comput. Sci. & Technol.,2002,17:549-555.
[19]	BREIMAN L. Random forests[J]. Machine Learning,2001,45(1):5-32. doi: 10.1023/A:1010933404324
[20]	TAALAB K,CHENG T,ZHANG Y. Mapping landslide susceptibility and types using random forest[J]. Big Earth Data,2018,2(2):159-178. doi: 10.1080/20964471.2018.1472392
[21]	TEKE A,KAVZOGLU T. Determination of effective predisposing factors using random forest-based Gini index in landslide susceptibility mapping[C]//Anon. 2nd Intercontinental Geoinformation Days (IGD). [S. 1.]:[S. n.],2021:198-201.
[22]	郭衍昊,窦杰,向子林,等. 基于优化负样本采样策略的梯度提升决策树与随机森林的汶川同震滑坡易发性评价[J]. 地质科技通报,2024,43(3):251-265. GUO Y H, DOU J, XIANG Z L, et al. Susceptibility evaluation of Wenchuan coseismic landslides by gradient boosting decision tree and random forest based on optimal negative sample samplingstrategies[J]. Bulletin of Geological Science and Technology,2024,43(3):251-265. (in Chinese with English abstract
[23]	CHEN T Q,GUESTRIN C. XGBoost:A scalable tree boosting system[C]//Anon. Proceedings of the 22nd Acm Sigkdd International Conference on Knowledge Discovery and Data Mining. San Francisco,CA,USA:[S. n.],2016:785-794.
[24]	FRIEDMAN J H. Stochastic gradient boosting[J]. Comput. Stat. Data Anal.,2022,38:367-378.
[25]	FAN Z Z,XU Y,ZHANG D. Local linear discriminant analysis framework using sample neighbors[J]. IEEE Transactions on Neural Networks,2011,22(7):1119-1132. doi: 10.1109/TNN.2011.2152852
[26]	付争方,朱虹,薛杉,等. 基于 Sigmoid 函数拟合的多曝光图像直接融合算法[J]. 仪器仪表学报,2015,36(10):2321-2329. doi: 10.3969/j.issn.0254-3087.2015.10.021 FU Q F,ZHU H,XUE S,et al. A direct fusion algorithm for multi-exposure images based on Sigmoid function fitting[J]. Journal of Instrumentation,2015,36(10):2321-2329. (in Chinese with English abstract doi: 10.3969/j.issn.0254-3087.2015.10.021
[27]	何平,刘紫燕. 基于改进多层感知机的手写数字识别[J]. 通信技术,2018,51(9):2075-2080. doi: 10.3969/j.issn.1002-0802.2018.09.011 HE P,LIU Z Y. Handwritten digit recognition based on improved multilayer perceptron[J]. Communication Technology,2018,51(9):2075-2080. (in Chinese with English abstract doi: 10.3969/j.issn.1002-0802.2018.09.011
[28]	CANTARINO I,CARRION M A,GOERLICH F,et al. A ROC analysis-based classification method for landslide susceptibility maps[J]. Landslides,2019,16:265-282 .
[29]	ZHENG X X,HE G J,WANG S S,et al. Comparison of Machine learning methods for potential active landslide hazards identification with multi-source data[J]. ISPRS Int. J. Geo-Inf.,2021,10:253. doi: 10.3390/ijgi10040253
[30]	杨城,林广发,张明锋,等. 基于 DEM 的福建省土质滑坡敏感性评价[J]. 地球信息科学学报,2016,18(12):1624-1633. YANG C,LIN G F,ZHANG M F,et al. DEM-based sensitivity evaluation of earthen landslides in Fujian Province[J]. Journal of Geo-Information Science,2016,18(12):1624-1633. (in Chinese with English abstract
[31]	李郎平,兰恒星,郭长宝,等. 基于改进频率比法的川藏铁路沿线及邻区地质灾害易发性分区评价[J]. 现代地质,2017,31(5):911-929. doi: 10.3969/j.issn.1000-8527.2017.05.004 LI L P,LAN H X,GUO C B,et al. Zonal evaluation of geologic hazard susceptibility along the Sichuan-Tibet Railway and adjacent areas based on improved frequency ratio method[J]. Modern Geology,2017,31(5):911-929. (in Chinese with English abstract doi: 10.3969/j.issn.1000-8527.2017.05.004
[32]	吴辰文,梁靖涵,王伟,等. 基于递归特征消除方法的随机森林算法[J]. 统计与决策,2017(21):60-63. WU C W,LIANG J H,WANG W,et al. Random forest algorithm based on recursive feature elimination method[J]. Statistics and Decision Making,2017(21):60-63. (in Chinese with English abstract
[33]	李萍,叶辉,谈树成. 基于层次分析法的永德县地质灾害易发性评价[J]. 水土保持研究,2021,28(5):394-399. LI P,YE H,TAN S C. Evaluation of geologic hazard susceptibility in Yongde County based on hierarchical analysis[J]. Soil and Water Conservation Research,2021,28(5):394-399. (in Chinese with English abstract
[34]	黄发明,李金凤,王俊宇,等. 考虑线状环境因子适宜性和不同机器学习模型的滑坡易发性预测建模规律[J]. 地质科技通报,2022,41(2):44-59. HUANG F M,LI J F,WANG J Y,et al. Modeling laws for landslide susceptibility prediction considering the suitability of linear environmental factors and different machine learning models[J]. Bulletin of Geological Science and Technology,2022,41(2):44-59. (in Chinese with English abstract
[35]	刘帅,朱杰勇,杨得虎,等. 不同降雨工况条件下的崩滑地质灾害危险性评价[J]. 地质科技通报,2024,43(2):253-267. LIU S,ZHU J Y,YANG D H,et al. Geological hazard assessment of collapse and landslide under different rainfall conditions[J]. Bulletin of Geological Science and Technology,2024,43(2):253-267. (in Chinese with English abstract