基于信息量和多层感知机分类器模型耦合的平果市斜坡类地质灾害易发性评价

王新伟; 张漓黎; 莫德科; 叶宗达; 江凡

doi:10.11932/karst20230208

基于信息量和多层感知机分类器模型耦合的平果市斜坡类地质灾害易发性评价

doi: 10.11932/karst20230208

王新伟^{1, 2,},
张漓黎³,
莫德科⁴,
叶宗达²,
江凡²

1.
中国有色桂林矿产地质研究院, 广西桂林 541000
2.
自然资源南方石山地区矿山地质环境修复工程技术创新中心, 广西南宁 530028
3.
广西壮族自治区桂林水文工程地质勘察院有限公司,广西桂林 541000
4.
广西壮族自治区二七四地质队, 广西北海 536005

基金项目: 桂林市2018年市本级可持续发展重大专项项目“典型废弃采石场生态修复及综合利用关键技术集成与示范”（合同编号：20180101-2）；国家重点研发计划课题“漓江流域喀斯特自然景观修复与植被生态功能提升关键技术研发及试验示范（合同编号：2019YFC0507503）”专题“漓江流域典型喀斯特废弃采石场景观修复技术与示范”

详细信息

作者简介:
王新伟（1988－），男，工程师，硕士，研究方向为水工环地质及地质环境修复治理。E-mail：wangxinwei1012@163.com

中图分类号: P245.25
计量
- 文章访问数: 782
- HTML浏览量: 464
- PDF下载量: 56
- 被引次数: 0
出版历程
- 收稿日期: 2022-07-01
- 刊出日期: 2023-04-25

Hillslope geo-hazard susceptibility assessment in Pingguo City based on coupling of CF information value and MLPC classifier model

1.
China Nonferrous Metals (Guilin) Geology and Mining Co., Ltd., Guilin, Guangxi 541000, China
2.
Technical Innovation Center for Restoration Engineering of Mine Geological Environment in Karst Areas of South China, Ministry of Natural Resources, Nanning, Guangxi 530028, China
3.
Guilin Survey Institute of Hydrogeological Engineering Co., Ltd., Guilin, Guangxi 541000, China
4.
No. 274 Geological Team of Guangxi, Beihai, Guangxi 536005, China

摘要

摘要: 广西平果市频发的地质灾害严重制约着市区的工程建设和生命财产安全。在充分收集和整理区域地质资料的基础上，通过遥感解译和现场调查，确定了平果市共发育251处斜坡类地质灾害，其中崩塌189处、滑坡62处。选择高程、坡度、坡向、曲率、工程地质岩组、距断层距离、土层厚度、距河流距离和降雨共9个因子作为评价因子，结合信息量和多层感知机分类器的优势，采用信息量和多层感知机分类器耦合模型对平果市斜坡类地质灾害进行易发性评价。斜坡类地质灾害易发性制图表明极高易发区占平果市面积的25.39%，主要分布于平果市的北部、中部和南部山区。通过ROC曲线对模型预测能力进行检验获得AUC=0.809，表明模型评价结果能够很好地预测研究区斜坡类地质灾害的发生。研究结果可为研究区的崩滑灾害风险评价和灾害防治提供科学依据。
- 平果市 /
- 斜坡类地质灾害 /
- 信息量 /
- 多层感知机分类器 /
- 易发性分区
Abstract: Frequent geological hazards in Pingguo City, Guangxi Province, seriously restrict the engineering construction in this area. Through the remote sensing interpretation and field investigation, 251 hillslope geological hazards, including 189 collapses and 62 landslides in Pingguo City are determined. All hazard points are randomly split into two subsets with 80% (200) and 20% (51) respectively for training and validation purposes. In addition, an equal number of non-hazard points were randomly selected from the hazard-free areas and then divided into a training dataset and a validation dataset. The hazard point is denoted as 1 and the non-hazard point is denoted as 0. Then, based on the geological environment conditions of Pingguo City, nine factors including elevation, slope angle, slope aspect, curvature, engineering geology groups, distance from faults, soil thickness, distance from rivers and rainfall are selected as hazard evaluation factors. Elevation, slope angle, slope aspect and curvature are topographic factors, which are produced from Advanced Land Observing Satellite (ALOS) digital surface model (DSM) with a resolution of 12.5 m. Distance from rivers is one of hydrological variables, also produced from the ALOS DSM. Engineering geology groups and distance from faults are geological factors, which are produced from a geological map at a scale of 1:200,000 provided by the China Geological Survey. Soil thickness is one of the factors related to land use, which is derived from domestic high resolutions remote sensing data such as Gaofen No.2 and Beijing No.2. Rainfall is an environmental factor, which is collected from the China Meteorological Administration. All relevant factors are converted into raster format with a 12.5-m resolution of ArcGIS software. Among these nine factors, engineering geology group is a categorical factor, whereas the other factors are continuous factors. In this study, continuous factors are reclassified into categorical factors based on natural breaks or equal interval methods. The tolerance (TOL) and variance inflation factor (VIF) are used to detect multicollinearity of all the factors. Generally, a VIF of less than 10 or a TOL of more than 0.1 indicates that all the factors are independent. Then, combined with the advantages of information value (IV) and multi-layer perceptron classifier (MLPC), the coupling model of IV and MLPC is used to evaluate the susceptibility of hillslope geo-hazard in Pingguo City, and the results are classified into five classes: very low, low, moderate, high, and very high. Finally, the receiver operating characteristic curve (ROC) and the area under the ROC curve are applied to carry out accuracy verification of the model.The main conclusions include: (i) Rainfall has the highest VIF value (1.163) and the lowest TOL value (0.86), both of which do not exceed the critical threshold values (TOL>0.1 or VIF<10). Thus, there are no severe collinearity problems among the 9 conditioning factors. (ii) The information values of the factor classes show that certain classes are conducive to the occurrence of hazards. The factor classes mostly prone to the occurrence of hazards are the elevation with 250-350 m, slope angle with 40°-50°, south slope, curvature less than -0.05, limestone and dolomite, 0-500 m from the faults, soil thickness with 0-1m, the area 0-250 m from the rivers and rainfall with 1,373-1,428 mm. (iii) The zoning map of hillslope geo-hazard susceptibility in Pingguo City shows that the areas with very high, high, moderate, low and very low susceptibility respectively account for 25.39%, 3.49%, 4.34%, 30.54% and 36.24% of the total area of the city. The areas with very high susceptibility are mainly distributed in the junction of steep mountains and plains, such as Liming town, Tonglao town and Bangwei town in the north of the study area, Haicheng town, Tonglao town and Pozao town in the middle and Xin'an town in the south. The hillslope geo-hazards in these five susceptible areas cover proportions of 64.54%, 8.37%, 1.59%, 21.91%, and 3.59%, respectively, which indicates that the evaluation results of hazard susceptibility are highly consistent with the distribution of historical collapse and landslides, and the mapping results of hazard susceptibility are reliable and accurate. (iv) The frequency ratio of each susceptibility zone can be obtained by calculating the ratio of hazard point proportion and area proportion of each susceptibility class. The frequency ratio increases from the very low susceptibility area to the very high susceptibility area, which indicates that the coupling model can effectively determine the different susceptibility levels in the study area. (v) The performance of the model is comprehensively compared through the AUC and ROC curves with validation datasets. The AUC value for the IV-MLPC model is 0.809, showing that the model exhibits a satisfactory performance (AUC>0.8) in collapse and landslide spatial prediction. The research results can provide a scientific basis for the risk assessment and disaster prevention of collapse and landslide in the study area.
- Pingguo City /
- hillslope geo-hazard /
- information value /
- multi-layer perceptron classifier /
- susceptibility zoning

HTML

图 1 研究区地理位置及斜坡类地质灾害编录

Figure 1. Location of the study area and hillslope geo-hazard inventory

下载: 全尺寸图片幻灯片

图 3 黎明乡那朗屯滑坡

Figure 3. Nalangtun landslide in Liming town

下载: 全尺寸图片幻灯片

图 2 龙板村崩塌

Figure 2. Collapse in Longban village

下载: 全尺寸图片幻灯片

图 4 灾害影响因子

Figure 4. Hazard influencing factors

下载: 全尺寸图片幻灯片

图 5 斜坡类地质灾害易发性分区图

Figure 5. Zoning of hillslope geo-hazard susceptibility

下载: 全尺寸图片幻灯片

图 6 IV-MLPC模型的ROC曲线图

Figure 6. ROC curve based on IV-MLPC model

下载: 全尺寸图片幻灯片

表 1 崩滑灾害影响因子分级及IV值

Table 1. The classes and IV of factors

因子	因子分级	分级栅格	分级栅格占比/%	灾害数目	灾害占比/%	IV
高程/m	50~150	1 801 053	11.324	15	5.976	−0.639
	150~250	3 231 404	20.318	60	23.904	0.163
	250~350	4 206 902	26.451	97	38.645	0.379
	350~450	3 916 670	24.627	61	24.303	−0.013
	450~550	1 995 484	12.547	14	5.578	−0.811
	550~650	649 804	4.086	4	1.594	−0.941
	650~896	102 919	0.647	0	0	0
坡度/°	0~10	4 306 296	27.076	2	0.797	−3.526
	10~20	4 354 216	27.378	8	3.187	−2.151
	20~30	3 869 997	24.333	24	9.562	−0.934
	30~40	2 037 339	12.81	46	18.327	0.358
	40~50	965 326	6.070	130	51.793	2.144
	50~60	311 711	1.960	36	14.343	1.990
	60~80	59 351	0.373	5	1.992	1.675
坡向	平面	206 476	1.298	0	0	0
	北	2 015 082	12.67	26	10.359	−0.201
	东北	1 863 358	11.716	26	10.359	−0.123
	东	1 820 559	11.447	37	14.741	0.253
	东南	2 250 940	14.153	30	11.952	−0.169
	南	2 189 502	13.767	46	18.327	0.286
	西南	2 002 428	12.591	36	14.343	0.130
	西	1 644 770	10.342	19	7.570	−0.312
	西北	1 911 121	12.016	31	12.351	0.027
曲率	−60.21~−0.05	6 661 594	41.886	131	52.191	0.220
	−0.05~0.05	3 046 564	19.156	52	20.717	0.078
	0.05~62.72	6 196 078	38.959	68	27.092	−0.363
工程地质岩组	岩组i	58 356	0.367	1	0.398	0.082
	岩组ii	5 033 581	31.649	63	25.100	−0.232
	岩组iii	3 273 833	20.585	34	13.546	−0.418
	岩组iv	7 538 466	47.399	153	60.956	0.252
距断层距离/m	0~500	3 722 914	23.408	66	57.769	0.903
	500~1 000	2 845 856	17.894	38	23.506	0.273
	1 000~1 500	2 159 709	13.579	29	14.343	0.055
	1 500~2 000	1 688 184	10.615	24	4.382	−0.885
	2 000~2 500	1 344 082	8.451	21	8.367	−0.010
	>2500	4 143 491	26.053	73	29.084	0.110
土层厚度/m	0~1	7 508 534	47.211	145	57.769	0.202
	1~3	4 051 848	25.477	59	23.506	−0.081
	3~5	3 647 726	22.936	36	14.343	−0.469
	＞5	696 128	4.377	11	4.382	0.001
距河流距离/m	0~250	5 831 693	36.668	152	60.558	0.502
	250~500	3 734 620	23.482	29	11.554	−0.709
	500~750	2 264 781	14.240	18	7.171	−0.686
	750~1 000	1 354 485	8.517	10	3.984	−0.760
	1 000~1 250	826 642	5.198	9	3.586	−0.371
	>1 250	1 892 015	11.896	33	13.147	0.100
降雨/mm	1 223~1 273	1 415 863	8.902	11	4.382	−0.709
	1 273~1 323	7 487 006	47.076	103	41.036	−0.137
	1 323~1 373	6 096 852	38.335	119	47.410	0.212
	1 373~1 428	904 515	5.687	18	7.171	0.232

下载: 导出CSV

表 2 地质灾害影响因子的共线性诊断

Table 2. Multicollinearity diagnosis of influencing factors of geo-hazards

影响因子	TOL	VIF
高程	0.937	1.067
坡度	0.895	1.117
坡向	0.980	1.021
曲率	0.964	1.038
工程地质岩组	0.855	1.169
距断层距离	0.987	1.013
土层厚度	0.960	1.042
距河流距离	0.896	1.116
降雨	0.860	1.163

下载: 导出CSV

表 3 研究区地质灾害易发区划统计

Table 3. Statistics of geo-hazard susceptibility zoning in the study area

易发性等级	灾害个数	灾害占比/%	面积/km²	面积占比/%	频率比	地质环境条件
极低易发区	9	3.59	900.58	36.24	0.10	地势相对平坦的区域，如高程50~150 m和坡度为0°~20°的地区
低易发区	55	21.91	758.82	30.54	0.72	50~150 m、坡度10°~30°、岩组iii
中易发区	4	1.59	107.82	4.34	0.37	高程150~250 m、坡度20°~40°、土层厚度d>5 m
高易发区	21	8.37	86.75	3.49	2.40	河谷两岸1 500 m范围内以及高程>450 m以及坡度30°~50°
极高易发区	162	64.54	631.05	25.39	2.54	高程150~350 m、坡度40°~60°、河流两岸500 m范围内，距离断层1 000 m范围内以及岩组（iv）

下载: 导出CSV

参考文献(24)

[1]	郑玲静, 李秀珍, 徐瑞池. 基于斜坡单元的区域滑坡敏感性评价: 以云南省小江流域为例[J]. 科学技术与工程, 2021, 21(28):12322-12329. doi: 10.3969/j.issn.1671-1815.2021.28.052 ZHENG Lingjing, LI Xiuzhen, XU Ruichi. Sensitivity evaluation of regional landslide based on slope units: Taking Xiaojiang river basin in Yunnan Province as an example[J]. Science Technology and Engineering, 2021, 21(28):12322-12329. doi: 10.3969/j.issn.1671-1815.2021.28.052
[2]	刘璐瑶, 高惠瑛, 李照. 基于CF与Logistic回归模型耦合的永嘉县滑坡易发性评价[J]. 中国海洋大学学报(自然科学版), 2021, 51(10):121-129. doi: 10.16441/j.cnki.hdxb.20200247 LIU Luyao, GAO Huiying, LI Zhao. Landslide susceptibility assessment based on coupling of CF model and logistic regression model in Yongjia county[J]. Periodical of Ocean University of China, 2021, 51(10):121-129. doi: 10.16441/j.cnki.hdxb.20200247
[3]	Huang Junpeng, Ling Sixiang, Wu Xiyong, Deng Rui. GIS-based comparative study of the Bayesian network, decision table, radial basis function network and stochastic gradient descent for the spatial prediction of landslide susceptibility[J]. Land, 2022, 11(3): 436.
[4]	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, 406. doi: 10.13869/j.cnki.rswc.2021.05.045 LI Ping, YE Hui, TAN Shucheng. Evaluation of geological hazards in Yongde county based on analytic hierarchy process[J]. Research of Soil and Water Conservation, 2021, 28(5):394-399, 406. doi: 10.13869/j.cnki.rswc.2021.05.045
[5]	吴远斌, 刘之葵, 殷仁朝, 雷明堂, 戴建玲, 罗伟权, 潘宗源. 基于AHP和GIS技术的湖南怀化地区岩溶塌陷易发性评价[J]. 中国岩溶, 2022, 41(1):21-33. WU Yuanbin, LIU Zhikui, YIN Renchao, LEI Mingtang, DAI Jianling, LUO Weiquan, PAN Zongyuan. Evaluation of karst collapse susceptibility in Huaihua area, Hunan Province based on AHP and GIS[J]. Carsologica Sinica, 2022, 41(1):21-33.
[6]	王念秦, 朱文博, 郭有金. 基于PSO-SVM模型的滑坡易发性评价[J]. 长江科学院院报, 2021, 38(4):56-62. doi: 10.11988/ckyyb.20200074 WANG Nianqin, ZHU Wenbo, GUO Youjin. Assessment of landslide susceptibility based on PSO-SVM model[J]. Journal of Yangtze River Scientific Research Institute, 2021, 38(4):56-62. doi: 10.11988/ckyyb.20200074
[7]	齐信, 黄波林, 刘广宁, 王世昌. 基于GIS技术和频率比模型的三峡地区秭归向斜盆地滑坡敏感性评价[J]. 地质力学学报, 2017, 23(1):97-104. doi: 10.3969/j.issn.1006-6616.2017.01.005 QI Xin, HUANG Bolin, LIU Guangning, WANG Shichang. Landslide susceptibility assessment in the Three Gorges area, China, Zigui synclinal basin, using GIS technology and frequency ratio model[J]. Journal of Geomechanics, 2017, 23(1):97-104. doi: 10.3969/j.issn.1006-6616.2017.01.005
[8]	杨根云, 周伟, 方教勇. 基于信息量模型和数据标准化的滑坡易发性评价[J]. 地球信息科学学报, 2018, 20(5):674-683. doi: 10.12082/dqxxkx.2018.170535 YANG Genyun, ZHOU Wei, FANG Jiaoyong. Assessment of landslide susceptibility based on information quantity model and data normalization[J]. Journal of Geo-information Science, 2018, 20(5):674-683. doi: 10.12082/dqxxkx.2018.170535
[9]	冯杭建, 周爱国, 唐小明, 游省易, 徐兴华. 基于确定性系数的降雨型滑坡影响因子敏感性分析[J]. 工程地质学报, 2017, 25(2):436-446. doi: 10.13544/j.cnki.jeg.2017.02.022 FENG Hangjian, ZHOU Aiguo, TANG Xiaoming, YOU Shengyi, XU Xinghua. Susceptibility analysis of factors controlling rainfall-triggered landslides using certainty factor method[J]. Journal of Engineering Geology, 2017, 25(2):436-446. doi: 10.13544/j.cnki.jeg.2017.02.022
[10]	杨永刚, 殷坤龙, 赵海燕, 黄晨忱, 陈丽霞, 张俞. 基于C5.0决策树-快速聚类模型的万州区库岸段乡镇滑坡易发性区划[J]. 地质科技情报, 2019, 38(6):189-197. doi: 10.19509/j.cnki.dzkq.2019.0622 YANG Yonggang, YIN Kunlong, ZHAO Haiyan, HUANG Chenchen, CHEN Lixia, ZHANG Yu. Landslide susceptibility evaluation for township units of bank section in Wanzhou district based on C5.0 decision tree and k-means cluster model[J]. Bulletin of Geological Science and Technology, 2019, 38(6):189-197. doi: 10.19509/j.cnki.dzkq.2019.0622
[11]	段宇英, 汤军, 刘远刚, 高贤君, 段宇雄. 基于随机森林的山西省柳林县黄土滑坡空间敏感性评价[J]. 地理科学, 2022, 42(2):343-351. doi: 10.13249/j.cnki.sgs.2022.02.016 DUAN Yuying, TANG Jun, LIU Yuangang, GAO Xianjun, DUAN Yuxiong. Spatial sensitivity evaluation of loess landslide in Liulin county, Shanxi based on random forest[J]. Scientia Geographica Sinica, 2022, 42(2):343-351. doi: 10.13249/j.cnki.sgs.2022.02.016
[12]	黄立鑫, 郝君明, 李旺平, 周兆叶, 贾佩钱. 基于RBF神经网络-信息量耦合模型的滑坡易发性评价: 以甘肃岷县为例[J]. 中国地质灾害与防治学报, 2021, 32(6):116-126. HUANG Lixin, HAO Junming, LI Wangping, ZHOU Zhaoye, JIA Peiqian. Landslide susceptibility assessment by the coupling method of RBF neural network and information value: A case study in Min Xian, Gansu Province[J]. The Chinese Journal of Geological Hazard and Control, 2021, 32(6):116-126.
[13]	王卫东, 刘攀, 龚陆. 基于支持向量机模型的四川省滑坡灾害易发性区划[J]. 铁道科学与工程学报, 2019, 16(5):1194-1200. doi: 10.19713/j.cnki.43-1423/u.2019.05.011 WANG Weidong, LIU Pan, GONG Lu. Landslide susceptibility mapping of Sichuan Province based on support vector machine[J]. Journal of Railway Science and Engineering, 2019, 16(5):1194-1200. doi: 10.19713/j.cnki.43-1423/u.2019.05.011
[14]	秦红富, 谈树成, 施旖奇, 李红梅, 汪旭. 基于CF-LR组合模型的地质灾害易发性评价: 以云南省宁洱哈尼族彝族自治县为例[J]. 人民长江, 2022, 53(7):119-127. QIN Hongfu, TAN Shucheng, SHI Yiqi, LI Hongmei, WANG Xu. Geological hazard susceptibility assessment based on CF&LR combined model: Case of Ning'er Hani and Yi Autonomous County, Yunnan Province[J]. Yangtze River, 2022, 53(7):119-127.
[15]	赵晓燕, 李永平, 谈树成. GIS支持下CF与信息量耦合模型的攀枝花市矿山地质灾害易发性评价[J]. 云南大学学报(自然科学版), 2022, 44(4):754-764. ZHAO Xiaoyan, LI Yongping, TAN Shucheng. Evaluation of mine geological hazard susceptibility of coupling CF with information model based on GIS in Panzhihua City[J]. Journal of Yunnan University (Natural Sciences Edition), 2022, 44(4):754-764.
[16]	李文彦, 王喜乐. 频率比与信息量模型在黄土沟壑区滑坡易发性评价中的应用与比较[J]. 自然灾害学报, 2020, 29(4):213-220. doi: 10.13577/j.jnd.2020.0422 LI Wenyan, WANG Xile. Application and comparison of frequency ratio and information value model for evaluating landslide susceptibility of loess gully region[J]. Journal of Natural Disasters, 2020, 29(4):213-220. doi: 10.13577/j.jnd.2020.0422
[17]	王志恒, 胡卓玮, 赵文吉, 国巧真, 万诗敏. 基于多层感知器模型的区域滑坡敏感性评价研究: 以四川低山丘陵区为例[J]. 防灾减灾工程学报, 2015, 35(5):691-698. WANG Zhiheng, HU Zhuowei, ZHAO Wenji, GUO Qiaozhen, WAN Shimin. Research on regional landslide susceptibility assessment based on multiple layer perceptron: Taking the hilly area in Sichuan as example[J]. Journal of Disaster Prevention and Mitigation Engineering, 2015, 35(5):691-698.
[18]	李晓青, 阳倩妮, 周楷淳, 罗为群. 喀斯特地区不同岩性上农村居民点分布特征: 以平果市为例[J]. 中国岩溶, 2021, 40(2):355-362. LI Xiaoqing, YANG Qianni, ZHOU Kaichun, LUO Weiqun. Distribution characteristics of rural settlement on different lithology in karst area: A case study of Pingguo City[J]. Carsologica Sinica, 2021, 40(2):355-362.
[19]	吴远斌, 刘之葵, 殷仁朝, 杨建兴, 罗伟权, 雷明堂, 戴建玲, 潘宗源. 湖南怀化盆地岩溶发育特征与分布规律[J]. 中国岩溶, 2022, 41(5):759-772, 807. WU Yuanbin, LIU Zhikui, YIN Renchao, YANG Jianxing, LUO Weiquan, LEI Mingtang, DAI Jianling, PAN Zongyuan. Karst development characteristics and distribution law in Huaihua basin, Hunan Province[J]. Carsologica Sinica, 2022, 41(5):759-772, 807.
[20]	董晓超, 熊康宁, 朱大运, 蓝家程, 廖建军, 曹洋, 刘兴宜. 喀斯特生态环境风险评价: 以贵州省为例[J]. 中国岩溶, 2019, 38(5):713-721. DONG Xiaochao, XIONG Kangning, ZHU Dayun, LAN Jiacheng, LIAO Jianjun, CAO Yang, LIU Xingyi. Risk assessment of karst ecological environments: A case study of Guizhou Province[J]. Carsologica Sinica, 2019, 38(5):713-721.
[21]	李艳慧. 广西平果至那坡地质灾害高易发时空分布及成因调查[J]. 赤峰学院学报(自然科学版), 2016, 32(21):41-43. doi: 10.13398/j.cnki.issn1673-260x.2016.21.018 LI Yanhui. Investigation on the spatial and temporal distribution and causes of geological disasters from Pingguo to Napo, Guangxi[J]. Journal of Chifeng University (Natural Science Edition), 2016, 32(21):41-43. doi: 10.13398/j.cnki.issn1673-260x.2016.21.018
[22]	Paola Reichenbach, Mauro Rossi, Bruce D. Malamud, Monika Mihir, Fausto Guzzetti. A review of statistically-based landslide susceptibility models[J]. Earth Science Reviews, 2018, 180(1):60-91. doi: 10.1016/j.earscirev.2018.03.001
[23]	李齐荣, 苗则朗, 陈帅, 李珂, 蒲明辉. 基于多模态分类模型的芦山地震滑坡易发性评价[J]. 长江科学院院报, 2022, 39(5): 63-70. LI Qirong, MIAO Zelang, CHEN Shuai, LI Ke, PU Minghui. Mapping the susceptibility of Lushan seismic landslide based on multimodal classification model[J]. Journal of Yangtze River Scientific Research Institute, 2022, 39(5): 63-70.
[24]	杜国梁, 杨志华, 袁颖, 任三绍, 任涛. 基于逻辑回归—信息量的川藏交通廊道滑坡易发性评价[J]. 水文地质工程地质, 2021, 48(5):102-111. doi: 10.16030/j.cnki.issn.1000-3665.202104009 DU Guoliang, YANG Zhihua, YUAN Ying, REN Sanshao, REN Tao. Landslide susceptibility mapping in the Sichuan-Tibet traffic corridor using logistic regression-information value method[J]. Hydrogeology & Engineering Geology, 2021, 48(5):102-111. doi: 10.16030/j.cnki.issn.1000-3665.202104009