吉首大学学报：自然科学版
吉首大學學報：自然科學版
길수대학학보：자연과학판
Journal of Jishou University(Natural Science Edition)
2012年
2期
80-83
,共4页
NLMS算法%DSP%Builder%自适应滤波器%FPGA
NLMS算法%DSP%Builder%自適應濾波器%FPGA
NLMS산법%DSP%Builder%자괄응려파기%FPGA
wavelet transform%image de-noise%threshold%threshold function
归一化LMS（NLMS）算法是一种变步长的LMS算法,比LMS算法具有更快的收敛速度.采用Altera公司的仿真软件DSP Builder和QUARTUSII7.2,进行归一化LMS算法的自适应滤波器的现场可编程门阵列（FPGA）设计.
歸一化LMS（NLMS）算法是一種變步長的LMS算法,比LMS算法具有更快的收斂速度.採用Altera公司的倣真軟件DSP Builder和QUARTUSII7.2,進行歸一化LMS算法的自適應濾波器的現場可編程門陣列（FPGA）設計.
귀일화LMS（NLMS）산법시일충변보장적LMS산법,비LMS산법구유경쾌적수렴속도.채용Altera공사적방진연건DSP Builder화QUARTUSII7.2,진행귀일화LMS산법적자괄응려파기적현장가편정문진렬（FPGA）설계.
The image denoising experiment, using biorthogonal wavelet and applying VisuShrink thresholds and BayesShrink thresholds,is conducted to images mixed with Gaussian white noise through the hard threshold function and the soft threshold function respectively. The results show that denoising effect of the hard VisuShrink threshold is better than that of soft VisuShrink threshold~ the denoising effect of the soft BayesShrink threshold is better than that of hard BayesShrink~and BayesShrink can have a better denoising effect only in the soft threshold function.
