基于边缘信息改进的双线性插值算法  

Improved bilinear interpolation algorithm based on edge information

作　　者：郭海霞[1] 郭海龙[2] 解凯[3] 

机构地区：[1]哈尔滨金融高等专科学校计算机系,哈尔滨150030 [2]哈尔滨学院,哈尔滨150080 [3]北京印刷学院信息与机电学院,北京102600

出　　处：《计算机工程与应用》2011年第31期171-174,共4页Computer Engineering and Applications

摘　　要：针对传统双线性插值算法插值时未考虑图像边缘信息,致使放大后图像出现模糊效应的特点,提出基于边缘信息的插值算法。该算法利用边缘检测及自适应门限划分出原始图像中的"边缘点"及"非边缘点",引入"灰度补偿"参数增加了边缘邻接插值点与非边缘邻接插值点间的灰度差,利用边缘点的"边缘方向"特性,判定待插值点是否位于放大图像的边缘上,对位于边缘上的待插值点利用边缘信息插值,恢复了使用传统双线性插值时断裂的边缘。大量的仿真实验证明该算法在一定程度上抑制了放大后图像边缘模糊的效应,改善了放大图像的视觉效果。Because edge information is not considered,some fringe details of image are damaged in the process of standard bilinear interpolation.To avoid the disadvantages of standard bilinear interpolation,a new algorithm based on edge information is proposed.In this method,three improved solutions have been proposed：edge-pixel and non edge-pixel are distinguished by edge-detection and adaptive-threshold,grey-difference between edge-adjoined-pixel and non edge-adjoined-pixel is extended.Using direction of edges,interpolated points are determined on edge or not.To the point on edge of magnified image,it is interpolated based on edge information.Many experiments confirm that this algorithm is superior to others.

关 键 词：双线性插值 图像处理 边缘检测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]