准确估计模态参数在飞机颤振边界的预测中具有重要意义。为了提高频域辨识算法的辨识效果 ,提出了一种用于颤振飞行试验数据处理的小波去噪方法。该方法引入梯度倒数加权滤波器对数据进行预处理 ,处理后的数据运用冗余小波进行小波分解 ,然后对输入信号在不同尺度下分别进行阈值降噪 ,对输出信号则采用了一种改进的小波空域相关滤波法去噪。最后通过仿真计算和实际数据证明该方法有效 Accurate estimation of modal parameters is of great importance in the prediction of aircraft flutter boundaries. In order to improve the identification effect of frequency domain identification algorithm, a wavelet denoising method for data processing of flutter flight test is proposed. In this method, the gradient inverse weighted filter is introduced to preprocess the data. The processed data is decomposed by the redundant wavelet, and then the threshold is de-noised separately for the input signal at different scales. An improved Wavelet airspace correlation filtering method denoising Finally, the simulation and actual data prove that the method is effective