引用本文:
【打印本页】   【HTML】   【下载PDF全文】   查看/发表评论  【EndNote】   【RefMan】   【BibTex】
←前一篇|后一篇→ 过刊浏览    高级检索
本文已被:浏览 13次   下载 7 本文二维码信息
码上扫一扫!
分享到: 微信 更多
山区公路沿线斜坡物理韧性随机森林评价模型——以四川省茂县为例
薛蒙蒙1, 文海家1,2,3, 林渝1, 孙德亮4
1.重庆大学 土木工程学院, 重庆 400045;2.山地城镇建设新技术教育部重点实验室, 重庆 400045;3.库区环境地质灾害防治国家地方联合工程研究中心, 重庆 400045;4.重庆师范大学 GIS应用研究重庆市高校重点实验室, 重庆 401331
摘要:
[目的] 建立山区公路沿线斜坡物理韧性随机森林评价模型,为山区斜坡抗灾工作提供科学参考。[方法] 以四川省茂县为研究区,选取高程、坡度、坡向、坡位、微地貌、曲率、顺逆向坡、归一化植被指数、岩性、距水系距离、距断层距离、距道路距离、多年平均降雨13个斜坡物理韧性评价因子,结合公路沿线498个历史斜坡破坏点,构建斜坡物理韧性评价的地理空间信息数据库。将样本数据按照7∶3的比例分为训练数据和验证数据。采用随机森林方法对训练数据进行训练建模,将得到的模型分别对训练数据、验证数据和样本整体数据进行预测分析。采用混淆矩阵和ROC曲线对模型预测的准确度进行验证。[结果] 评价因子中高程、距道路距离、距水系距离这3个评价因子的权重较大。该模型精度较高,混淆矩阵的精度为98.9%,训练数据集、验证数据集和整个研究区的ROC曲线下的面积(AUC)值分别为1.000,0.870和0.978。模型仿真到整个研究区中,将研究区的物理韧性划分为极低、低、中、高、极高5个等级。[结论] 基于随机森林方法构建的山区公路沿线斜坡物理韧性评价模型具有较高的稳定性以及可靠性。
关键词:  山区公路沿线斜坡  物理韧性  随机森林模型  地理空间数据
DOI:10.13961/j.cnki.stbctb.2020.04.023
分类号:U41;X43;X9
基金项目:重庆市技术创新与应用示范一般项目(社会民生类)“山地建筑韧性评价大数据分析建模与应用示范”(cstc2018jscx-msybX0310)
Random Forest Evaluation Model for Physical Toughness of Slopes Along Mountain Roads -Taking Maoxian County of Sichuan Province as an example
Xue Mengmeng1, Wen Haijia1,2,3, Lin Yu1, Sun Deliang4
1.School of Civil Engineering, Chongqing University, Chongqing 400045, China;2.Key Laboratory of New Technology for Construction of Cities in Mountain Area of the Ministry of Education, Chongqing 400045, China;3.National Joint Engineering Research Center of Geohazards Prevention in the Reservoir Area, Chongqing 400045, China;4.The Key Laboratory of GIS Application Research, Chongqing Normal University, Chongqing 401331, China
Abstract:
[Objective] A random forest evaluation model for physical toughness of slopes along mountain roads was established in order to provide a scientific reference for disaster prevention in mountain areas. [Methods] Taking Maoxian County, Sichuan Province as the research area, this study selected 13 physical toughness assessment factors of slopes, including elevation, aspect, slope direction, slope position, micro landform, curvature, type of slope, normalized vegetation index, lithology, distance from water system, distance from fault, distance from road and annual average rainfall, and combined with 498 historical slope failure points along the highway, to construct a geospatial information database for the evaluation of slope physical toughness. The sample data was divided into training data and validation data according to the proportion of 7∶3. The random forest method was used to train and model the training data, then the obtained model was used to predict and analyze the training data, the validation data and the overall sample data respectively. And confusion matrix and ROC curve were used to verify the accuracy of the model prediction. [Results] Among the evaluation factors, the weight of elevation, distance from the road, and distance from the water system was higher. The accuracy of the model was high, the accuracy of confusion matrix was 98.9%, and the AUC (area under the ROC curve) values of the training data, the validation data and the overall sample data was 1.000,0.870 and 0.978, respectively. The model was simulated into the entire study area, and the physical toughness of the study area was divided into 5 levels as: extremely low, low, medium, high and extremely high. [Conclusions] The physical toughness evaluation model of the slope along the mountain highway based on the random forest method has high stability and reliability.
Key words:  slopes along mountain roads  physical toughness  stochastic forest model  geospatial data