摘要
针对Mallat算法构造小波变换矩阵处理图像.为解决有限域脊波变换(FRIT)只能应用于素数像素大小的图像且图像处理中产生随机信号和"环绕"的问题,提出3种改进型FRIT方法,并将其与不同边界延拓方法相结合应用于图像去噪中.实验结果表明,所给方法在对于具有直线特征的图像去噪方面性能较优.
The authors constructed wavelet transformation matrix to process images in view of Mallat algorithm,in order to solve the questions that FRIT can only apply to pixel-size image of prime number,generated random signal and ＂around＂ phenomenon in image processing,and proposed three kinds of improved FRIT methods that were combined with boundary prolong methods,which were used in the image denoising,the experiment showed that the denoising performance of the image with a straight line feature is better.
出处
《吉林大学学报（理学版）》
CAS
CSCD
北大核心
2010年第1期99-103,共5页
Journal of Jilin University:Science Edition
基金
国家科技支撑计划项目基金(批准号:2007BAE17B04)
