基于Kalman滤波融合算法的库区滑坡动态变形监测综合信息提取  

APPLICATION OF MULTI-SENSOR VALUATION IN DYNAMIC DEFORMATION FOR LANDSLIDE MONITORING BASED ON KALMAN FILTRATION FUSION THEORY

作　　者：彭鹏[1] 单治钢[1] 董育烦[1] 陈长河[1] 付钦祯 

机构地区：[1]中国电建华东勘测设计研究院,浙江杭州310014

出　　处：《岩石力学与工程学报》2014年第8期1520-1525,共6页Chinese Journal of Rock Mechanics and Engineering

基　　金：国家自然科学基金资助项目(50579089)

摘　　要：为了消除和减小滑坡变形多点监测中的误差,正确评价滑坡体的工程地质特性及其时效性,采用基于Kalman滤波的决策级融合方法对西南某滑坡变形进行动态分析。在融合监测年间的水平位移和垂直位移数据的基础上,综合判断滑坡体的依时性工程特性。融合结果表明该滑坡体先后经历缓慢变形期、匀速变形期、加速变形期和急剧变形期,表明该滑坡体具有阶段性变化的特点,且降雨入渗是诱发该滑坡发生位移变形的主要因素,结果符合滑坡体变形特征的自然规律。同时修正了传统方法不考虑传感器自身因素的局限性,合理利用融合方法的互补性,消除了部分传感器采集数据的模糊性、不确定性和随机性,证明该方法在滑坡动态变形监测与分析中具有有效性和可行性。The method of decision-level fusion based on Kalman filteration was used in order to reduce and even eliminate the errors in multi-sensor monitoring and to correctly evaluate the geological characteristics of a landslide in Southwest China. The data of horizontal and vertical displacements were analysed. Results show that the landslide had experienced stages of slow deformation,uniform deformation,accelerated deformation and rapid deformation. It indicates that the landslide is a stepped process of variation. The rainfall infiltration is the main factor. The Kalman filtration fusion theory reduces the limitations of traditional methods by making use of the fusion method. It eliminated the ambiguity of the data collected by sensors. The feasibility and effectiveness of the method are confirmed.

关 键 词：边坡工程 多传感器系统 KALMAN滤波 数据级融合方法 动态变形监测 

分 类 号：P642[天文地球—工程地质学]