In sensor performance test

signal of sampling is often subjected to interference and pollution of all kinds of noise

which can not accurately reflect operating status of equipment

and should not be directly used for data processing and analysis. In order to carry out de-noise preprocessing on test datas

according to characteristics of specific sensor test data

using wavelet transform method

preprocessing wavelet determine approperate wavelet denoising parameters. Introduce reconstruction factor

compare several kinds of commonly used wavelet basis

and then decomposition level is determined according to smoothness

simulation signal is processed by wavelet threshold de-noising use Matlab

which provide an effective wavelet threshold de-noising method for character test data preprocessing of sensor.