摘要
论文在分析了硬阈值算法的优点与不足之处的基础上,提出了一种新的硬阈值图像去噪算法,并将此算法成功运用到临场感系统中。该算法不仅继承了硬阈值算法处理误差较小、图像边缘保持良好和实现简单的优点,而且针对原始算法阈值划分粗糙、光滑度损失等缺点,引入了调整因子和相邻窗口等方法降低阈值估计风险和提高图像光滑度。理论证明及实验结果表明,新算法是一种高效的实时去噪算法,其不仅具有良好的视觉效果和较低计算复杂度,而且在相同噪声情况下,新算法的MSE、PSNR等指标均优于硬阈值算法。
In this paper,a new hard threshold algorithm is proposed after analyzing the advantages and disadvantages of the traditional hard threshold algorithm,and it applies successfully on the tele-presence system.The new algorithm not only includes the advantage of the traditional in which less error,good image edge,and that is implemented easily,but also reduce the risk of threshold estimation and improve the image smoothness,using the methods of adjusting factor and neighbourhood window.The results of the analysis and experiment demonstrate that the new algorithm is a effective real time de-noising method,not only has the properties of good vision effect and less complexity,but also performs better than that of traditional algorithm in the aspects of MSE,PSNR etc under the basis of the same noise.
出处
《计算机工程与应用》
CSCD
北大核心
2005年第33期83-85,94,共4页
Computer Engineering and Applications
关键词
临场感
小波
闽值硬阚值算法
tele-presence,wavelet,threshold,hard threshold algorithm
