一种基于双边滤波和梯度信息的滤波算法  

A Filtering Algorithm Based on Bilateral Filtering and Gradient Information

作　　者：范娟 陈昌海 王静 李振锋 FAN Juan;CHEN Changhai;WANG Jing;LI Zhenfeng(Sichuan province equipment manufacturing industry robot application Technology Engineering Laboratory,Deyang 618000,China)

机构地区：[1]四川省装备制造业机器人应用技术工程实验室,四川德阳618000

出　　处：《电视技术》2018年第11期7-10,共4页Video Engineering

摘　　要：针对双边滤波算法会使弱边缘变得模糊和不易硬件实现这一问题,文章提出了一种基于双边滤波和梯度信息的改进滤波算法。该算法加入了基于梯度信息的梯度相似度因子,能够在滤波过程中加强对弱边缘的保护,使得滤波后图像的边缘更加清晰。为了能够较小算法的时间开销,使用N阶多项式替换指数项的方法,文章提出了快速改进滤波算法。仿真结果表明,改进滤波算法具有比双边滤波算法更好的滤波效果,弱边缘处的畸变得到了极大的改善;当合理选择参数N时,快速改进滤波算法能够极大的缩小时间开销,同时具有不弱于双边滤波算法的滤波效果。To solve the problems that the traditional bilateral filter algorithm would blur the weak edges and is difficult to implement with hardware,an improved filtering algorithm based on bilateral filtering and gradient information is proposed in this paper.The gradient similarity factor based on gradient information is considered,which can strengthen the protection of weak edges in the filtering process,so that the edge of the filtered image is more clear.In order to reduce the time cost of the improved algorithm,a fast improved filtering algorithm is proposed by replacing the exponential term with the Nth-order polynomial.The simulation results show that,the distortion at the weak edge has been greatly improved for that the improved filtering algorithm has better filtering effect than the bilateral filtering algorithm,and the fast improved filtering algorithm can greatly reduce the time cost and have better filtering result than the bilateral filtering algorithm when the parameter N is reasonably selected.

关 键 词：双边滤波 梯度信息 时间开销 滤波算法 

分 类 号：TP393.17[自动化与计算机技术—计算机应用技术]