摘要
为了有效去除X射线数字底片的噪声,本文研究和建立了X射线数字底片的噪声模型,在传统的小波萎缩阈值去噪算法基础上,提出一种改进型的阈值去噪算法。该算法对小波变换的各系数进行Bayes参数估计,并根据子图直方图对噪声进行参数估计从而进行阈值设置,阈值处理后通过小波逆变换还原图像。实验结果表明,该算法同传统图像去噪算法相比,图像处理后较光滑清晰,视觉效果好,具有较高的峰值信噪比,能满足X射线数字底片后续处理的要求。
In this paper, a modeling of X-ray digital film image noise has been constructed. An improved threshold denoising method was presented based on the traditional wavelet threshold method. The new method estimated the wavelet coefficient using Bayes rule. It set threshold value for more times for the sake of better view. It simulated two methods to de-noise the image, one was traditional wavelet transform, and the other was the improved method. Simulation result shows that higher peak-signal-to-noise （PSNR） ratio and better view can be obtained by using the later comparing with the anterior image de-noising methods. It can meet the request of the subsequent X-ray digital film process better.
出处
《电子测量技术》
2007年第2期33-35,95,共4页
Electronic Measurement Technology
基金
国家核能开发项目(B030206)资助
关键词
X射线数字底片
噪声模型
小波变换
变阈值去噪
X-ray digital negative
modeling of noise
wavelet transform
changing threshold
de-noising
