基于SVM的泥石流危险度评价研究
收稿日期: 2007-04-04
修回日期: 2007-08-11
网络出版日期: 2008-03-20
基金资助
南京邮电大学"攀登计划"项目(NY206075)资助。
Debris Flow Hazard Assessment Based on SVM
Received date: 2007-04-04
Revised date: 2007-08-11
Online published: 2008-03-20
选取泥石流一次(可能)最大冲出量(L1)、泥石流发生频率(L2)、流域面积(S1)、主沟长度(S2)、流域最大相对高差(S3)、流域切割密度(S6)和泥沙补给段长度比(S9)7个因子作为泥石流沟谷危险度评价因子,运用支持向量机理论,以云南省37条泥石流沟的259个基础数据为样本进行学习训练和测试,建立泥石流危险度评价的支持向量机模型,通过实例验证,取得良好效果。
原立峰 . 基于SVM的泥石流危险度评价研究[J]. 地理科学, 2008 , 28(2) : 296 -300 . DOI: 10.13249/j.cnki.sgs.2008.02.296
In order to improve the limitation of traditional debris flow hazard assessment methods, a SVM-based debris flow hazard assessment method was proposed. Seven factors including the most volume of once flow (L1), frequency (L2), watershed area (S1), valley length (S2), watershed relative height difference (S3), valley incision density(S6) and the length ratio of sediment supplement (S9) were chosen as assessment factors of debris flow hazard degree. Using support vector machine (SVM) theory, selecting Radial Basis Function, and using trial-and-error method for optimal selection of parameters, C=8, r=2. Thirty seven debris flow channels with 259 basic data in Yunnan Province were selected as training samples, and an assessment model based on SVM was created. The model was applied to evaluating debris flow hazard degree of Jishi Valley hydropower station of Huanghe (Yellow) River. Assessment result consistency came to 73.33% comparing to fuzzy mathematic method. The results show that the model has advantage of best generation, high training speed, and convenient for modeling through an instance application. It will be thought as being broad application scope that SVM was applied to hazard assessment of debris flow.
[1] 钟敦伦, 谢洪, 王士革,等. 北京山区泥石流[M]. 北京: 商务印书馆, 2004:175~176.
[2] 汤家法, 谢洪. GIS技术支持下的泥石流危险度区划研究: 以岷江上游为例[J]. 四川测绘, 1999,3:120~122.
[3] 朱静, 唐川. 云南省泥石流灾害危险度分区研究[J]. 中国地质灾害与防治学报, 1996,7(2):86~93.
[4] 胡凯衡,李泳,韦方强.泥石流流域集水区面积限值与一级水系数目关系[J].地理科学,2005,25(4):473~477.
[5] 陈杰,韦方强,崔鹏.小江流域泥石流堆积扇形成的制约因素及其特征[J].地理科学,2005,25(6): 704~701.
[6] 陈宁生,张飞.2003年中国西南山区典型灾害性暴雨泥石流运动堆积特征[J].地理科学,2006,26(6):701~711.
[7] 闫满存,王光谦.基于GIS的澜沧江下游区滑坡灾害危险性分析[J].地理科学,2007,27(3):365~370.
[8] 刘家宏, 王光谦. 基于遥感图像的泥石流地面活动程度评价[J]. 地理科学, 2003, 23(4):454~459.
[9] 唐川, 张军, 万石云,等.基于高分辩率遥感影像的城市泥石流灾害损失评估[J].地理科学,2006,26(3): 358~363.
[10] 唐川, 朱大奎. 基于GIS技术的泥石流风险评价研究[J]. 地理科学,2002,22(3):300~305.
[11] 闫满存,王光谦,刘家宏. GIS支持的澜沧江下游区泥石流爆发危险性评价[J]. 地理科学, 2001, 21 (4) 334~338.
[12] 许兴旺. 怀柔县泥石流沟谷危险度FUZZY评判[J]. 安庆师范学院学报, 1997,3(4):14~17.
[13] 于秀治, 韦京莲. 灰色系统理论在北京山区泥石流危险度评价预测中的应用[J].中国地质灾害与防治学报, 2004,5(1):118~120.
[14] 汪明武, 金菊良, 李丽. 投影寻踪新方法在泥石流危险度评价中的应用[J]. 水土保持学报, 2002,16(6):79~81.
[15] 刘涌江, 胡厚田, 白志勇. 泥石流危险度评价的神经网络法[J]. 地质与勘探, 2001,37(2):84~87.
[16] 马志江, 陈汉林, 杨树锋. 基于支持向量机理论的滑坡灾害预测——以浙江庆元地区为例[J]. 浙江大学学报, 2003,30(5):592~596.
[17] Christopher J C Burges. A Tutorial on Support Vector Machine for Pattern Recognition[J]. Data Mining and Knowledge Discovery, 1998, 2: 121-167.
[18] 李国正, 王猛, 曾华军(译). 支持向量机[M]. 北京: 电子工业出版社, 2004:18.
[19] 刘希林, 唐川. 泥石流危险性评价[M]. 北京: 科学出版社, 1995:15~26.
[20] 邓乃扬, 田英杰. 数据挖掘中的新方法——支持向量机[M]. 北京: 科学出版社, 2004.
[21] Liu X. Assessment on The Severity of Debris Flows in Mountainous Creeks of Southwest China.Proceedings of International Symposium of Interpraevent. Germany:Carmisch-Partenkirenchen, 1996: 145-154.
[22] 胡亚东, 傅荣华. 夏克勤. 黄河积石峡水电站库区泥石流危险度评价[J]. 灾害学, 2004, 19(2):36~41.
/
| 〈 |
|
〉 |