基于控制函数估计图像噪音的标准方差.pdfVIP

Vo l. 36 No.9 第 36 卷第9 期 计算机科学 2009 年 9 月 Computer Science Sep 2009 慕于控制函数估计图像噪音的标准方提 王文远 (北京应用物理与计算数学研究所 北京 100094) 摘 要提出了一种新的估计自然阁像噪音的标准方巍的算法。该算法通过控制品数在经过网格化后的图像上选择 某些区域，然后用一种梯皮滤波算法去估计噪音的标准方恙。控制函数是一个与阁像的预信噪比相关的囊，可以周总 体误萃的最小化来优化确定。控制函数可以有效地平衡阁像结构对噪音估计的影响，因而不像已有的算法只能对图 像有限吃围的噪音强度进行有效估计。所提算法能够对从极小到极大把阁的噪音强度进行准确估计。同时，整个噪 音估计算法可以通过一个快速收敛的选代算法来计算，以期获得更稳健的解。量化实验表明，相对于已有的算法，新 算法仿计图像噪音的标准方j且是准确。 关键诩 标准方差，网格化，控制函数，预信噪比 中阔法分提号 TP39 1. 41 文献栋iR码 A Estimating the Standard Deviation of Noise via ControlJed Function WANG Wen-yuan Onstitute of Applied Physics and Computational Mathematics ,Beijíng 100094 ,China) Abstract This paper presented a new algorithm to estimate the standard deviation of noise for nature images. It usωa gradient filter to estimate the noise on some parts which are selected by a controlled function from the tessellat吨ima­ ge. The controlled function can be related to the preestimate of SNR We optimally determined the controlled function as an exponential function by minimizing the total en侃The control1ed function can effectively offset the effect wlùch is caused by filtering the image structur已 Therefore unlike some existing algorithms which can only accurately estimate noise for images with limited range of noise intensity ,the propos时 algorithm can perfectly do this for images with the noise intensity range of extremely smal1 to extremely h创vy. Moreover. an iterative process which has fast convergence was implemented to obtain more robust solution. Quantitative ωmparisons with some existing algorith