超限邻域平均算法在肺部CT图像除噪中的应用

Application of Lungs CT Image Denoising with Overlimit Neighborhood Average Algorithm

本文针对含有高密度椒盐和高斯噪声的肺部CT图像除噪中细节信息保留不够,图像较模糊、清晰度欠佳等问题,提出了一种超限邻域平均算法,该方法首先通过阈值策略法对指定邻域内像素加权均值与其中任一像素灰度值大小进行比较判断,然后将大于阈值的像素剔除,而小的作为有用信息输出,最后运用该方法对含有高密度椒盐和高斯噪声的医学影像图像进行除噪实现设计。仿真实验表明,超限邻域平均算法对肺部CT图像的高密度椒盐和高斯噪声抑制力强,计算速度快,峰值信噪比(PSNR)比单纯中值和均值算法大,而均方根误差(RMSE)小,且去噪后的图像具有良好的细节保真度和清晰度。

To deal with the problem of medical CT images with high density noise, poor definition, an over-limit neighborhood averaging algorithm was proposed.The method compared the weighted mean of designated pixel with any pixel gray value by threshold strategy, then removed the pixels larger than the threshold, took the small output as useful information, and finally designed the medical imagimg with high-density salt and pepper noise, and Gaussian noise.Simulation experiments showed that the over-limit neighborhood average algorithm had strong high density noise salt and pepper, and Gaussian noise suppression on lung CT images, fast computation. Peak signal to noise ratio(PSNR) was larger than pure median and mean algorithm, smaller root mean square error(RMSE), and the denoising image had good detail fidelity and clarity.