基于分数阶微分的滑坡变形监测数据的融合

Fusion of Landslide Deformation Monitoring Data Based on Fractional Differential

摘要: 目前滑坡监测方式主要是点监测，但这种监测手段只能反映滑坡局部的变形情况，不能完全反映滑坡面上整体的变形情况，而且在测量过程中，监测值往往会受到随机噪声的干扰，影响各个传感器测量的结果，这两种情况都对滑坡的预测预报有极大的影响。本文将分数阶微积分与数据融合两种理论相互结合，建立分数阶微分滑坡数据的融合处理模型，通过该算法得到一个更全面的信息，然后用滑坡实际监测信息进行实验计算。结果表明：分数阶微分具有显著的数据融合效果，与动态权值加权相比，该算法精度更高的同时能更好地反馈滑坡的全局信息，有利于提高滑坡监测的准确性和预测的科学性。

Abstract: At present, the main monitoring method of landslide is point monitoring, but this monitoring method can only reflects the local deformation of landslide, not the whole deformation on landslide surface, and in the process of measurement, the monitoring value is often disturbed by random noise and affects the measurement results of each sensor. Both cases have great influence on landslide prediction. In this paper, two theories of fractional calculus and data fusion are combined to establish a fusion processing model of fractional differential landslide data. A more comprehensive information is obtained by this algorithm. Then the actual monitoring information of landslide is used for experimental calculation. The results show that fractional differential has significant data fusion effect. Compared with dynamic weight weighting, the algorithm has higher accuracy and better feedback of global information of landslide. It is helpful to improve the accuracy of landslide monitoring and the scientific nature of prediction.