【摘要】：針對地震滑坡規(guī)律復雜,預測難度大,而傳統(tǒng)地震滑坡敏感性分析采樣策略研究不足的問題,分析不同采樣策略對地震滑坡敏感性分析的影響。以多邊形緩沖區(qū)作為采樣區(qū),并將滑坡的發(fā)生區(qū)域及前兩者總和2個采樣策略用于對比實驗。選取4·20蘆山地震中受災嚴重的寶盛鄉(xiāng)作為研究對象,根據(jù)采樣策略對滑坡各個評價因子進行采樣,并構(gòu)建支持向量機模型定量計算敏感性指數(shù),再利用自然斷點法生成敏感性區(qū)劃圖。通過多種方法對結(jié)果進行分析。其中,以多邊形緩沖區(qū)為采樣策略得到的滑坡正確率94.44%,受試者工作特征曲線下面積99.1%,種子單元面積指數(shù)綜合評價效果最佳,3項結(jié)果均占優(yōu)勢。結(jié)果證明該采樣策略是有效可行的,可為后續(xù)防災減災工作提供依據(jù)。
[Abstract]:Aiming at the problem that the law of earthquake landslide is complex and the prediction is difficult, and the traditional sampling strategy of seismic landslide sensitivity analysis is insufficient, the influence of different sampling strategy on seismic landslide sensitivity analysis is analyzed. The polygonal buffer zone is used as the sampling area, and the two sampling strategies of the occurrence area and the sum of the first two are used in the comparison experiment. Baosheng Township, which was badly affected by the Lushan earthquake, was selected as the research object. According to the sampling strategy, the evaluation factors of landslide were sampled, and the support vector machine model was constructed to quantitatively calculate the sensitivity index. Then the natural breakpoint method is used to generate the sensitivity map. The results are analyzed by various methods. Among them, the correct rate of landslide obtained with polygonal buffer as sampling strategy is 94.44, the area under the operating characteristic curve is 99.1 and the comprehensive evaluation effect of seed unit area index is the best, and the three results are superior. The results show that the sampling strategy is effective and feasible, and can provide the basis for the subsequent disaster prevention and mitigation work.
【作者單位】： 中國地質(zhì)大學地球物理與空間信息學院;武漢工程大學資源與土木工程學院;中國地震局地震研究所;
【基金】：國家“863”高技術(shù)研究發(fā)展計劃資助項目(2012AA121303)~~
【分類號】：P642.22

【相似文獻】

相關期刊論文 前10條

1 陳曉利;王U，

本文編號：2311287