摘要
图像配准是超分辨率图像重建的关键.提出了一种基于R滤波器的超分辨率图像重建的三步算法.第一步:对原图进行加窗模糊;第二步:根据全局运动模型配准,计算平移参数和旋转角度,对图像进行平移和旋转;第三步:把序列低分辨率图像配准到同一个坐标系中,重构高分辨率图像.实验表明,三步算法的配准精度和超分辨率图像重建效果较好.
Image registration is the key to super-resolution image reconstruction.A three-step registration algorithm for super-resolution images based on R filter is presented.The first step is to obfuscate the original image by adding windows.The second step is to revolve and translate the image by calculating translation parameters and rotation angle according to the requirement of the global motion model.The last step is to match image sequences of low resolution in the same coordinate system and to reconstruct the high resolution images.Experiment results show that matching accuracy and result of reconstructing super-resolution images are greatly improved.
出处
《内蒙古大学学报（自然科学版）》
CAS
CSCD
北大核心
2010年第4期443-449,共7页
Journal of Inner Mongolia University：Natural Science Edition
基金
黑龙江省教育厅科学技术项目(编号11531388)资助
关键词
图像配准
图像重建
超分辨率
三步法
image registration image reconstruction super-resolution three-step method
