Kalman滤波是一种处理动态数据的有效的方法,其在大坝变形数据处理中将会有很好的应用前景。但当存在异常时,Kalman滤波精度将会降低甚至滤波发散。BP神经网络是一个高度复杂的非线性动力学系统,它可以处理含有不确定因素的问题,可以任意精度逼近。因此BP神经网络辅助Kalman滤波能在一定程度上提高滤波的精度。 Kalman filtering is an effective method to deal with dynamic data. It will have a good application prospect in dam deformation data processing. However, when there is anomaly, the Kalman filtering accuracy will be reduced or even the filtering will diverge. BP neural network is a highly complex nonlinear dynamics system that can deal with problems with uncertainties and can be approximated with any precision. Therefore, BP neural network aided Kalman filter to a certain extent, improve the filtering accuracy.