论文部分内容阅读
传统的实时信号处理消噪方法常采用FFT ,但用其处理非平稳信号或奇异信号时 ,效果并不明显 .在文中一方面基于信号、噪声在多尺度空间能量分布的差异而推出的Donoho非线性子波阈值消噪原理 ,另一方面从子波变换与FIR滤波器组的特定关系出发 ,在滤波卷积、下二采样过程中 ,将数据按一定规律重排 ,利用可编程逻辑器件CPLD设计了一个对非平稳信号进行实时子波分解及阈值消噪的专用芯片 .
The traditional real-time signal processing denoising methods often use FFT, but the effect is not obvious when dealing with non-stationary signals or singular signals.On the one hand, Donoho based on the difference of signal and noise energy distribution in multi-scale space Linear wavelet threshold denoising principle, on the other hand, starting from the wavelet transform and the specific relationship between the FIR filter group, the filter convolution, the next two sampling process, the data rearranged according to certain rules, the use of programmable logic device CPLD A special chip for real-time wavelet decomposition and threshold denoising of non-stationary signals is designed.