基于图像自相似结构的视频图像插值算法研究  

Research on video image interpolation algorithm based on image self-similar structure

作　　者：向豪 孙冬[1] 高清维[1] 卢一相[1] XIANG Hao;SUN Dong;GAO Qingwei;LU Yixiang(School of Electrical Engineering and Automation,Anhui University,Hefei 230601,China)

机构地区：[1]安徽大学电气工程与自动化学院,安徽合肥230601

出　　处：《传感器与微系统》2021年第5期56-59,63,共5页Transducer and Microsystem Technologies

基　　金：安徽省教育厅高校科学研究项目(KJ2019A0023,KJ2019A0022)。

摘　　要：针对在处理边缘复杂和纹理信息丰富的图像时,传统的视频图像插值算法难以达到较高的清晰度,输出图像容易产生锯齿和模糊效应的问题,提出了一种新的视频图像插值算法。算法利用帧间和帧内图像大量存在的空间冗余信息,建立了空域分形自相似模型。该模型在非局部区域搜索得到邻居块,通过对低分辨图像块和不同采样网格上的邻居块线性组合来拼贴重建得到高分辨率图像。实验结果表明:该算法插值后的图像重建精度较高,同时由于模型的结构较为简单,整个算法的计算量也很小。Aiming at the problem that traditional video image interpolation algorithm is difficult to achieve high definition when dealing with complex and texture-rich images,and the output image is prone to aliasing and blurring effects,a video image interpolation algorithm is proposed;it utilizes inter-frame and intra-frame spatial redundancy information,which searches for non-local regions to obtain neighbor patches,and reconstructs high-resolution(HR)by collaging low-resolution(LR)patches and linear combinations of neighbor patches on different sampling grids.The experimental results show that the image reconstruction accuracy of the algorithm is high,and the calculation of the whole algorithm is small due to the simple structure of model.

关 键 词：图像插值 空间冗余 非局部自相似 采样网格 图像块拼贴 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]