论文部分内容阅读
选择地貌特征(包括地形坡度、沟谷切割状况、洼地封闭状况、前缘临空状况)、滑面特征(包括滑面倾角、滑面形态、滑面阻抗比)、滑体结构和近期活动迹象4大类9个因子作为古(老)滑坡类潜在滑坡的判识指标,以三峡库区和岷江上游地区相同数量的典型滑坡作为学习样本,依据距离判别分析(DDA)和Fisher线性判别分析(FLDA)方法,分别建立了两个研究区潜在滑坡的DDA判识模型和FLDA判识模型。实例分析结果表明,DDA方法对两个研究区学习样本和测试样本的误判率均为0,而FLDA方法对两个研究区测试样本的误判率分别为5.56%和11.11%,对学习样本的错判率分别为11.1%和0%。因此,在三峡库区和岷江上游地区潜在滑坡的判识中,DDA法比FLDA法判识准确性更高、适用性更强。
Geomorphological features (including topographic gradient, ditch cutting conditions, depression closure conditions, leading edge heading conditions), slip surface characteristics (including slip surface inclination, slip surface morphology, slip surface impedance ratio), slip structure and recent activity 4 According to the distance discriminant analysis (DDA) and the Fisher linear discriminant analysis (FLDA), the nine kinds of factors are used as the identification index of the potential landslides in the ancient (old) landslide. Based on the same number of typical landslides in the Three Gorges reservoir area and the upper reaches of the Minjiang River as learning samples, ) Method, the DDA discriminant model and the FLDA discriminant model of potential landslides in the two study areas are respectively established. The results of case analysis show that the false positive rate of DDA is 0 for both learning and test samples in the two study areas, while the false positive rate of FLDA is 5.56% and 11.11% for the two experimental areas respectively. The miscarriage rates were 11.1% and 0% respectively. Therefore, in the identification of potential landslides in the Three Gorges reservoir area and the upper reaches of Minjiang River, the DDA method has higher accuracy and applicability than the FLDA method.