摘要
双边滤波是一种保边去噪的滤波算法,但原始双边滤波算法的执行时间长,且滤波精确度差,不能有效去除‘椒盐噪声’。利用像素的空间位置特征和灰度分布特征,在三维空间中缩短双边滤波的执行时间,保证滤波精确性。首先,像素灰度作为第三维,根据灰度参数和空间参数对图像进行划分,形成许多体素。然后,选择每个体素的中心作为基本点,计算每个体素的滤波值。最后利用像素的三维坐标,确定像素点所在体素,并通过线性插值计算每个像素的滤波值,输出滤波图像。在实验中分别对具有"椒盐噪声"的不同格式图像进行滤波。实验结果表明改进的算法执行时间短,并能有效地滤除图像中的"椒盐"噪声,保证滤波精确性。
Bilateral filtering is an algoritlim to filter noises while preserving image edges.A naive implementation time of tlie bilateral filtering is long and the filtering precision is poor.It can+1 effectively filter the‘salt and pepper'noise.In order to shorten the executing time of bilateral filtering and guarantee the accuracyof filtering,the location feature and gray distributionfeature of pixels in 3Dspaceare used.At first,pixel gray is taken as thethirddimension.Accorciingto the gay parameter and space parameter,the imageis dividedand formmany voxels.Then takethe center of voxel as basic point and computethefilteringvaliue of every voxel.At last,according to3'coordinates ofpixel,identify where pixel is in voxel and uselinear interpolation tocompute fil-tering value of pixels andoutput filtering image.In experiment,“salt and pepper”noise in different format images are filtered.Experiment results showthat improved algorithmhas a short executing time.It effectively filters the‘salt and pepper'noise in im-age and guarantees the filtering precision.
作者
郑丽娟
张杰
郑丽萍
ZHENG Lijuan;ZHANG Jie;ZHENG Liping(School of Business,Liaocheng University,Liaocheng 252000,China;School of Computer Science,Liaocheng University,252000,China)
出处
《电视技术》
2018年第12期11-17,共7页
Video Engineering
基金
国家自然科学基金资助项目(71701085)
国家教育部人文社会科学研究青年基金项目(16YJCZH159)
山东省高等学校人文社会科学研究项目(J16YF25)
聊城大学科研项目(31801140)
关键词
双边滤波
体素
基本点
线性插值
椒盐噪声
Bilateral Filtering
Voxel
Basic point
Linear Interpolation
Salt and pepper noise
