一种基于二进小波变换改进图像的去噪方法  

Image Restoration Method Based on B-spline Binary Wavelet Transform

作　　者：马瑞瑞 王刚[1] 张静[1] MA Ruirui;WANG Gang;ZHANG Jing(School of Mathematical Sciences, Xinjiang Normal University, Urumqi 830017, China)

机构地区：[1]新疆师范大学数学科学学院,乌鲁木齐830017

出　　处：《渭南师范学院学报》2022年第5期87-92,共6页Journal of Weinan Normal University

基　　金：新疆师范大学优秀青年教师科研启动基金资助项目:三维八向双正交小波包及其性质(XJNU202014)。

摘　　要：传统的图像去噪方法会造成图像细节的丢失,且去噪效果一般,结合二进小波变换、中值滤波和图像相加运算的优点,提出一种基于B-样条二进小波变换的图像恢复方法。利用二维中值滤波和图像的加法运算对方差为0.005的高斯噪声图像复原,然后利用B-样条二进小波滤波器将噪声图像分解一次再重构,得到复原后图像,对上述三种方法实验结果做了对比:原图像已知时,图像的加法运算、去噪效果最佳;原图像未知时,B-样条二进小波滤波效果最佳,去噪效果优于中值滤波的去噪效果。然后,提出新的算法,用B-样条二进小波重构后的图像分别进行二维中值滤波去噪和图像的加法运算去噪。结合的方法相较于单一使用的方法有一定提升,并且保留了更多的图像细节信息。The traditional image denoising method will cause the loss of image details,and the denoising effect is general.Combining the advantages of binary wavelet transform,median filtering and image addition,an image based on B-spline binary wavelet transform is proposed.The Gaussian noise image with a variance of 0.005 is restored by two-dimensional median filtering and image addition,and then the noise image is decomposed and reconstructed once by using the B-spline binary wavelet filter to obtain the restored image.The experimental results of the method are compared.When the original image is known,the image addition operation has the best denoising effect;when the original image is unknown,the B-spline binary wavelet filter has the best effect,and the denoising effect is better than the median filter.In the second step of the experiment,a new algorithm is proposed.The image reconstructed by B-spline binary wavelet is used for two-dimensional median filter denoising and image addition denoising respectively.The following conclusions are drawn:the combined method has a certain improvement compared with the single method,and retains more image details.

关 键 词：B-样条二进小波变换 中值滤波 图像的加法运算 图像复原 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]