基于Keren改进配准算法的IBP超分率重建  

IBP Super-resolution Reconstruction Based on Improvement Approach of Keren Registration Method

作　　者：张永育[1] 李翠华[1] 余礼钹[1] 张东晓[1] 李雄宗[1] 施华[1] 

机构地区：[1]厦门大学信息科学与技术学院,福建厦门361005

出　　处：《厦门大学学报（自然科学版）》2012年第4期686-690,共5页Journal of Xiamen University：Natural Science

基　　金：国防基础科研计划项目;国防科技重点实验室基金项目;高等学校博士学科点专项科研基金项目(20110121110020)

摘　　要：提出了一种基于Keren改进配准算法的迭代反投影(iterative back-projection,IBP)超分辨率重建算法.该算法克服了Keren迭代配准算法基于小角度旋转的局限,并在迭代运算过程中引入了权重因子和阈值.权重因子有效地控制了算法的收敛速度,提高算法的稳定性.阈值的引入使得算法效率更高,配准结果更加准确.通过Keren改进配准算法进行配准,再通过IBP算法对配准后图像序列进行超分辨率重建,仿真结果表明,基于Keren改进配准算法的IBP重建具有良好的超分辨率重建效果.This paper proposed an iterative back-projection（IBP） super-resolution reconstruction algorithm based on Keren improve-ment registration method. Keren improved algorithm combines the ideas that overcame Keren algorithm based on small angle rota tion. And in the iteration process this algorithm have introduced weighting factors and threshold. Weighting factor effectively control the convergence rate and improve the algorithm stability. The introduction of the threshold makes the algorithm more efficient, the registration results more accurate. By Keren improvement registration method for alignment,and then through the IBP after registra tion algorithm for super resolution reconstruction of image sequences. Simulation results show that the IBP super resolution recon- struction based on the Keren improved registration algorithm have good results. It has a good application value.

关 键 词：超分辨率 图像配准 迭代反投影(IBP) 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]