结合全变分的双边滤波图像去噪方法研究

摘要

The image denoising,as a key step in digital image processing,affects the final quality of image processing di-rectly. Since the traditional total variation(TV)regularization denoising algorithm is easily to produce the staircase effect,a bi-lateral filtering method based on TV model is proposed to eliminate the staircase effect to a certain extent,in which the filtering characteristics in the aspects of spatial domain and value domain are considered. The experimental results show that the denoising method can obtain better denoising effect,reserve the edge feature information of the image,and has perfect denoising effect. In comparison with PSNR obtained with TV algorithm,the PSNR of the gray image with 256 × 256 size is increased by 1.45 dB, and the PSNR of the gray image with 512×512 size is increased by 2.56 dB.%图像去噪是数字图像处理过程中的一个重要步骤,它将直接影响到图像处理的最终质量.针对传统的全变分(TV)正则化去噪算法容易产生阶梯效应的缺点,利用双边滤波去噪算法在空间域和值域两个方面进行滤波的特点,提出一种结合TV模型的双边滤波方法,该方法能在一定程度上有效地改善阶梯效应.仿真实验结果表明,提出的去噪方法不仅能够获得较好的去噪效果,还能有效地保持图像的边缘特征信息,降噪效果明显.在较高水平噪声情况下,与TV算法相比,该方法针对小尺寸灰度图片(256×256)图像的峰值信噪比(PSNR)提高1.45 dB左右,大尺寸灰度图片(512×512)图像的PSNR提高2.56 dB左右.