基于区域分割的自适应反锐化掩模算法  

Adaptive unsharp masking method based on region segmentation

作　　者：杨词银[1] 尚海波[2] 贾晨光[2] 黄廉卿[1] 

机构地区：[1]中国科学院长春光学精密机械与物理研究所,吉林长春130022 [2]海军装备部沈阳军代表局,辽宁沈阳110031

出　　处：《光学精密工程》2003年第2期188-192,共5页Optics and Precision Engineering

基　　金：国家自然科学基金资助项目(No.69677015)

摘　　要：针对线性反锐化掩模法的缺点,提出一种基于区域分割的自适应反锐化掩模法,利用像素局部方差将输入图像划分为平滑区、中等对比度区和高对比度区,根据像素(x,y)所属的区域类型,自适应确定其局部灰阶变化率增益系数α(x,y)及期望的输出局部灰阶变化率Hd(x,y),并导出增强系数K(x,y)。对一幅胸片图像进行处理,并与其它几种反锐化掩模法的处理结果进行比较,结果显示该方法获得较好的边缘增强效果,且有较强的噪声抑制能力。A adaptive unsharp masking method based on region segmentation is presented aiming at the defects of linear unsharp masking. The input image is divided into homogeneous areas, medium contrast areas and large contrast areas by the local variance of image pixels. Base on the area type to which image pixel (x,y) belongs, the local activity gain factor α(x,y) and the desired output local activity Hd(x,y) are determined adaptively, and the enhancement factor K(x,y) is derived. A X_ray chest image is processed with the proposed method and other unsharp masking methods. The results show that good capabilities of edge enhancing and noise suppressing are achieved by the proposed method.

关 键 词：自适应反锐化掩模 局部灰阶变化率 边缘增强 噪声 算法 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]