自适应门限子波消噪的专用芯片研究

ASIC design for DWT de-noising based on adaptive soft-threshold

本文从子波变换与完全重构FIR滤波器组的特定关系出发 ,将输入数据按一定规律重新排列之后实时地将数据分解到不同的尺度空间 ,然后经过正交子波基的Donoho非线性软门限消噪处理。同时 ,推导出信号休止期内的相对实时自适应噪声方差算法。在此子波消噪专用芯片的研究过程中 ,利用可编程逻辑器件CPLD进行了时序仿真和验证 。

According to the relationship of wavelet transforms and perfect reconstructional FIR filter banks,this paper presents an algorithm for de noising with Donoho non linear soft threshold in defferent levels of multiscale space through rearranging the input data before convolving and filtering and sub sampling And give a simple iterative algorithm to caculate the variance of the noise in interregna It works well wherever the signal or noise is stationary or not At last,it is implemented with CPLD for the future VLSI design and real time applications