中文摘要：

对小波变换阈值函数算法进行了改进，并实现了高铁变形数据的去噪处理，使得去噪后均方根误差相对于改进前明显减小，信噪比相对于改进前明显增大，改进效果明显。算例表明，用Kalman滤波模型对改进阈值函数小波去噪结果进行预测，预测效果较好，说明本文改进方法是一种有效的去噪方法，可以用于建筑物变形分析预处理。

英文摘要：

An improved algorithm of wavelet transform threshold function is proposed,by the denoising process-ing of high-speed rail deformation data.Compared with previous methods,this algorithm reduces the denoising a lot.When RMSE is decreased,the signal-to-noise is obviously increased.A prediction is made for the im-proved threshold function wavelet denoising by Kalman filter model.This improved method is proved to be an effective denoising method and can be applied to the pretreatment of the building deformation analysis.