基于BP神经网络的区域滑坡易发性评价

Evaluation of Regional Landslide Susceptibility Assessment Based on BP Neural Network

区域滑坡易发性评价是开展区域滑坡地质灾害危险性、风险性评价的基础。结合新疆伊宁县野外地质调查数据,采用数据挖掘技术分析研究区黄土滑坡控制性因素,以此作为挑选致灾因子的判据。通过BP神经网络模型来构建区域滑坡易发性预测模型,采用训练良好的BP神经网络模型,结合整个研究区DEM数据和遥感解译数据,得出研究区滑坡灾害易发性分区图,为当地区域滑坡的预防和治理决策提供一定参考。

The evaluation of regional landslide susceptibility is the basis of regional landslide hazard assessment and regional landslide risk assessment. Combined with the field geological survey data of Yining County in Xinjiang, the controlling factors of landslide in the study area are analyzed by using data mining technology, which can be used as the criteria for selecting disaster causing factors. The BP neural network model is used to build the prediction model of regional landslide susceptibility. The trained BP neural network model is combined with the DEM data and remote sensing interpretation data of the whole study area to obtain the landslide hazard susceptibility zoning map of the study area, which provides a certain reference for the local regional landslide prevention and governance decision-making.