基于小波变换模极大值的去噪方法研究-河北工业科技-河北科技大学.PDFVIP

27 6 河 北 工 业 科 技 Vol. 27, No . 6 2010 11 H ebei Journal of Industrial Science and T echnology No . 20 10 : 2010) , , ( 河北科技大学信息科学与工程学院, 河北石家庄 0500 18) : 小波去噪在信号处理中得到广泛应用 前常用的方法主要有 Donoho 提出的阈值法和 Mallat 提出的模极大值法模极大值法是一种经典的小波去噪方法, 噪声的模极大值的幅度随尺 度的增大而迅速减小, 而正常信号随尺度的增大而增大, 因此利用合适尺度的小波变换, 容易把噪 声从正常信号中剔除通过试验说明这种方法对白噪声和脉冲噪声都有很好的去噪效果, 并与阈 值去噪比较, 对于高斯白噪声, 信噪比比较低的信号, 模极大值去噪要优于阈值法去噪; 对于脉冲噪 声, 脉冲噪声点数较多时, 模极大值去噪要优于阈值法去噪 : 噪声; 小波变换; 模极大值; 阈值去噪; 信噪比 : T N911. 7 : A Research in denoising method based on w a elet transform modulus max ima LIU Limei, LIU Qiyue, ZHANG Jing ( College of Information Science and Engineering , H ebei Uni ersity of Science and T echnology, Shijiazhuang H ebei 050018, China) Abstract: Wa elet denoising is w idely used in signal processing. The current methods commonly used are mainly the thresh old method proposed by Donoho and the modulus max ima proposed by Mallat. The modulus max ima method is a classical w a e let denoising method. The modulus max ima amplit ude of the noise decreases rapidly w ith the increase of scale, w hile the nor mal signals amplitude increases w ith the increase of scale. So appropriate scale of wa elet transform can easily remo e noise from the normal signal. T est pro es the method has ery good denoising effect on w hite noise and impulse noise. For Gaussian white noise, when SNR is low , modulus max ima denoising method is superior to threshold denoising ; for impulse noise, when there are many impulse points, the modulus maxima denoising is superior to threshold denoising. Key words: noise; w a elet t ransform; modulus max ima; threshold denoising ; SNR , , , , Donoho