一种新型大米检测图像的自适应滤波算法  

A New Adaptive Filter Algorithm of Rice-Detection Images

作　　者：冯月亮[1] 肖助明[1] 李涛[1] 

机构地区：[1]湖南科技大学信息与电气工程学院,湘潭411201

出　　处：《中国粮油学报》2006年第6期151-154,173,共5页Journal of the Chinese Cereals and Oils Association

摘　　要：针对大米检测图像中的脉冲噪声提出了一种新型的基于极值和噪声密度估计算法的自适应滤波算法,结合了EM算法和IMFLED算法的优点。实验表明该算法比现有的中值滤波去噪和保持物体的实际大小、边缘等许多重要细节的能力更强,特别在噪声密度比较高的情况。An adaptive median filter for Rice Detection System is presented, which is based on extreme value and estimate noise detection for the removal of impulse noise from rice images. It combines the merits of IMFLED algorithm （ a new effective filtering algorithm for the removal of impulse noise from Image） and EM algorithm （ a new filtering algorithm based on extriemmun and median）. Simulation results demonstrate that this algorithm is superior to other existing median filters in preserving the origin information such as primitive size and edge of object, being especially effective for the cases where the images are extremely highly corrupted.

关 键 词：图像处理 极值 噪声估计 自适应中值滤波 

分 类 号：TN713[电子电信—电路与系统]