论文部分内容阅读
为了有效地去除光纤陀螺信号中的分形噪声,提出了全频域小波多尺度阈值降噪方法。从分形信号的分数微分形式出发,建立了分形信号与高斯白噪声之间的联系,结合小波分析的高阶消失矩特性,实现了分形信号在小波空间的去相关。当分形信号被具有高阶消失矩的小波分解后,其在小波域具有白噪声的特点,可以采用阈值处理的方法予以去除。同时,根据陀螺信号噪声的宽频带特点,对小波分解的低频近似系数和高频细节系数进行阈值处理,有效地抑制了噪声成分。光纤陀螺信号的降噪实例表明:相比传统的小波阈值降噪方法、卡尔曼滤波方法和滑动平均滤波方法,该方法具有较好的降噪效果。
In order to effectively remove the fractal noise in the fiber optic gyroscope signal, a multiscale threshold denoising method in the whole frequency domain is proposed. Based on the fractal and differential forms of fractal signals, the relationship between fractal signals and Gaussian white noise is established. Combined with the high order vanishing moments of wavelet analysis, the fractal signal is decorrelated in the wavelet space. When the fractal signal is decomposed by the wavelet with high-order vanishing moments, it has the characteristic of white noise in the wavelet domain and can be removed by threshold processing. At the same time, according to the characteristics of the wideband noise of the gyro signal, the low frequency approximation coefficient and the high frequency detail coefficient of the wavelet decomposition are thresholded to effectively suppress the noise component. The example of noise reduction of fiber optic gyroscope shows that this method has a better noise reduction effect than the traditional wavelet threshold denoising method, Kalman filtering method and sliding average filtering method.