小波阈值去噪算法是去除图像信号中白噪声的有效算法。其中阈值选取关系到图像重建的精确度。软硬阈值函数能够有效地去除噪声,却不能很好地保留细节信息。在传统软硬阈值函数的基础之上提出了一种自适应的闽值函数,能根据分解层次的不同而自动地调节阈值。根据仿真结果,证明该算法进一步继承了硬软阈值函数的优点,改善了其缺点,而且能够较好去掉噪声并且保留图像的细节。
A wavelet threshold denoising algorithm is effective for denoising the white noise in image. The selection of threshold will affect the quality of the reconstruction image. The soft and hard thresholding function can get rid of the noise effectively, but it cant well reserve the detail information. Based on the soft and hard thresholding function, a thresholding function, which can adjust itself by the decomposed layers, is proposed. Experimental results show that this algorithm inherits the advantage of the hard and soft thresholding function and overcomes their weakness; moreover, it can well reserve the detail of the image.