基于三次B样条插值曲面的图像放大方法  

An Image Magnification Method Based on B-spline Interpolation Surfaces

作　　者：孙志中[1] 王磊[1] 桑伟泉 

机构地区：[1]贵州大学计算机科学与技术学院,贵阳550025

出　　处：《计算机与数字工程》2015年第3期477-479,515,共4页Computer & Digital Engineering

基　　金：国家自然科学基金(编号:61262006;60863005);贵州省科学技术基金(编号:20122125)资助

摘　　要：为了实现低分辨率图像的高保真放大,利用原图像中相邻四个像素点的像素值及其在放大后的图像中的位置构造出了对应的三次B样条插值曲面,从而较好地估算出了放大后图像中出现的空白像素值。鉴于彩色图的像素点有R、G、B三个分量,分别构造R曲面、G曲面、B曲面对空白像素点的三个分量进行填充。由于采用构造曲面的高度值作为像素值,使得相邻像素点的像素值不存在较大差异,故而尽可能地避免了马赛克现象的产生,实现了图像的高保真放大。通过对比实验,验证了论文方法对图像放大具有非常好的效果。In order to fulfill the high fidelity magnification of images,this paper used the pixel value of four adjacent pixel in the original image and their respective position after the image was enlarged to construct the B-spline interpolation surfaces,and the height of the surfaces were estimations of the pixel values.The pixel of colored images could be divided into three components as R,G and B.This paper did paddings for the vacant pixels of R surface,G surface,and B surface respectively.As the height of constructed surfaces were taken as the pixel values,there was not many differences among the pixel values of adjacent pixels,which safely avoided the phenomena of mosaic as soon as possible,and fulfilled the high fidelity magnification of images.The method of this paper had very good effect on the image magnification which was verified by experiment.

关 键 词：插值 B样条曲线 B样条曲面 图像放大 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]