摘要
依据白噪声小波变换性态与信号奇异性相比具有显著不同的特点，在大尺度下设置阈值，去掉噪声信号而保留图像细节信号引起的模极大值点。在阈值设置问题上，采用自适应阈值的方法，克服单一阈值不能在每级尺度上将信号与噪声作最大分离的缺点。实验表明，与单一阈值去噪方法相比，该方法不仅可以保留图像边缘信息，而且能提高去噪后图像的峰值信噪比2～5dB。
It has been proved that noise wavelet transform apparently differs from signal sigularty in that the modulus maximums caused by noise signal are removed and image details are preserved by setting thresholds in big scale. As for setting the threshold value, the application of the adaptive threshold value can overcome the weakness that the only one threshold value can not separate signal from noise in the each class dimensions. The experiment shows that this method can not only reserve the imagePs edge information, but also raise the imagers SNR in the top to 2-5 dB after being denoised.
出处
《海军工程大学学报》
CAS
北大核心
2007年第1期44-47,共4页
Journal of Naval University of Engineering
关键词
小波变换
模极大值
自适应阈值
去噪
wavelet transform
modulus maximum
adaptive threshold value
denoising
