多尺度形态学在彩色图像去噪中的应用  

Use of Multi-scale Morphology in Filtering out Noise of Color Image

作　　者：贾书香[1] 任小洪[1] 王天文[1] 梁立飞[1] 

机构地区：[1]四川理工学院自动化与电子信息学院,四川自贡643000

出　　处：《计算机技术与发展》2010年第1期128-131,共4页Computer Technology and Development

基　　金：四川省(高校)重点基金资助项目(2007R011)

摘　　要：文中阐述了多尺度形态学对彩色图像滤噪。该方法采用大小不同的结构元素进行图像滤波,通过大尺度抑制噪声,小尺度得到丰富的边缘。根据彩色图像各颜色通道RGB的相光敏感曲线,选择结构元素,对图像进行滤波去噪并进行合并。结果表明该方法对噪声有较好的抑制作用,保持了原图像的细节信息,具有一定的可行性和实用性。根据此方法的特点,该方法适用的图像类型是图像中的对象尺寸都比较大,没有细小的细节,对这种类型的图像除噪的效果会比较好。Multi-scale morphology is introduced into the color image for filtering out noise.Different sizes of structural elements are used for filtering out noise of image in the method.Then use anti-noise characteristics of large-scale to curb noise and use position properties of small-scale to get the rich edge.According to the relative spectral sensitivity curve of the RGB color channel of color image,structural elements are separately selected to filter out noise and merge.The results show that the method has better noise inhibition, to maintain the details of information of the original image and has feasibility and practicality. According to the characteristics of this method, the method is suitable for the type of images, which the size of the object in images is large and has no small detail. The effects of filtering out noise of image would be better.

关 键 词：多尺度形态学 结构元素 彩色图像 滤除噪声 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]