一种新的改进阈值的小波包图像去噪方法

A wavelet packets image de-noising method based on improved threshold

介绍了小波阈值去噪的基本原理,针对软、硬阈值化方法在图像处理上存在的边界模糊和振荡等失真的不理想情况,采用了一个新的阈值化构造函数。该函数运用了阶次调节因子和振荡调整因子,能分别对作用对象进行动态修正,以获得较优的小波系数的阈值估计。仿真试验表明,采用该阈值化函数的小波包消噪方法能对带噪图像进行有效去噪,对比实验统计表明,其性能优于传统的阈值化方法。

In this paper, the theory of wavelet threshold de-noising has introduced, and a new threshold function was presented, in order to overcome distortions of hard and soft threshold with oscillation and borderline blurred. Regulatory factor of Order and oscillation adjustment factor are important in this new function; they can be adjusted properly to produce the best estimations of the wavelet coefficients. The outcome of experiments indicates that the new wavelet packets threshold algorithm can de-noise efficiently and statistics shows the improved method can obtain the best de-noising effects by adjusting the parameter values compared with traditional threshold functions.