基于最优化彩色空间转换的多梯度插值算法  

The multi-color gradient image interpolation algorithm based on optimized color space transform

作　　者：李杰[1] 程磊[1,2] 吴怀宇[1] 陈洋[1] 

机构地区：[1]武汉科技大学信息科学与工程学院,湖北武汉430081 [2]北京大学工学院,北京100871

出　　处：《华中科技大学学报（自然科学版）》2013年第S1期371-374,共4页Journal of Huazhong University of Science and Technology(Natural Science Edition)

基　　金：国家自然科学基金资助项目(60705035;61075087;61203331);武汉科技大学冶金工业过程系统科学湖北省重点实验室开放基金重点资助项目(Z201102);河南省高等学校控制工程重点学科开放基金资助项目(KG2011-01);湖北省教育厅科研计划重点资助项目(D20131105);湖北省自然科学基金重点资助项目(2010CDA005)

摘　　要：针对彩色插值算法在处理图像方面存在图像边缘模糊现象的问题,提出一种基于最优化彩色空间转换的多梯度彩色插值算法.该算法首先利用多梯度算子来选择插值方向以恢复绿色分量;然后选择一个有效的准则函数去完成最优化彩色空间转换,从而恢复另外两个缺失的颜色分量.在胶囊机器人微视觉平台上,通过与传统的插值算法进行对比实验,结果显示:此算法优于传统的插值算法,在图像边缘和纹理方面有所提高,改善了图像的视觉效果,从而证明了其在胶囊内窥镜上应用的有效性.As for current color interpolation algorithms which may exist the problem of edge blurring in the aspects of image processing,a multi-color gradient image interpolation algorithm was proposed based on optimized color space transform.The algorithm firstly restores the green component by using the multi-color gradient to effectively select interpolation direction.Then,accomplish optimal color space transform by selecting a valid criterion function,so the other two missing components of color can be restored in an optimized color space.Finally,by comparing this algorithm to conventional interpolation algorithms on a capsule robot micro-vision platform,results demonstrate that this algorithm is superior to the traditional interpolation algorithm.The images edge and texture have been improved,and the visual effects of images have also been improved.Therefore,the effectiveness is obvious through the application of this algorithm on the capsule endoscope.

关 键 词：彩色插值算法 彩色空间转换 双线性插值 边缘检测 Bayer型彩色滤波阵列 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]