基于曲线拟合的自适应滤波算法  

ADAPTIVE FILTERING ALGORITHM BASED ON CURVE FITTING

作　　者：董蕊[1] 王志巍[2] 刘淑娟[2] 张有会[2] 

机构地区：[1]河北交通职业技术学院基础教学部河北石家庄050091 [2]河北师范大学数学与信息科学学院河北石家庄050016

出　　处：《计算机应用与软件》2013年第6期43-45,共3页Computer Applications and Software

基　　金：国家自然科学基金项目(10771049)

摘　　要：针对低密度脉冲噪声,提出基于曲线拟合的自适应滤波算法。首先结合以点邻域为判别区域的方法检测出噪声点;其次对噪声点,根据均方差最小化原则自适应的选择模板内全信号点方向;最后结合曲线拟合确定模板中心像素点的灰度值,实现被污染的数据点的有效恢复。实验表明,这种方法具有较好的细节保护能力和较强的噪声去除能力。Aiming at the low density impulse noise, we point out an adaptive filtering algorithm based on curve fitting. First, the noise points are detected by the means of making point neighbourhood as the differentiation region. Then, the whole signal point direction within the template is selected adaptively according to the mean variance minimisation principle. Finally, the contaminated data points are recovered effectively by identifying the gray value of the template centre pixel in combination with curve fitting. Experiment shows that this method has the ability of detail protection and noise removal.

关 键 词：曲线拟合 自适应 噪声检测 均方差 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]