虚拟现实场景中模糊图像盲复原算法研究  

Research on Blind Restoration Algorithm of Blurred Image in Virtual Reality Scene

作　　者：苏智[1] 熊兵[2] 陈岭[3] SU Zhi;XIONG Bing;CHEN Ling(Hunan Changsha University of Science&Technology,Changsha Hunan 410114,China;School of Computer Science and Communication Engineering,Changsha University of Science&Technology,Changsha Hunan 410114,China;College of engineering and design,Hunan Normal University,Changsha Hunan 410081,China)

机构地区：[1]长沙理工大学,湖南长沙410114 [2]长沙理工大学计算机与通信工程学院,湖南长沙410114 [3]湖南师范大学工程与设计学院,湖南长沙410081

出　　处：《计算机仿真》2023年第10期228-232,共5页Computer Simulation

基　　金：2022年度湖南省社会科学成果评审委员会研究课题(XSP22YBC174);2021年湖南省教育厅科学研究项目(编21C0191);2021年湖南省社会科学成果评审委员会课题(XSP21YBZ099)。

摘　　要：在模糊图像盲复原处理中,如果忽略图像像素的权重,会导致复原后图像的结构相似度较低,于是提出虚拟现实场景中模糊图像盲复原算法。引入多尺度方法对目标图像进行分解,并估算每一层图像的模糊核。将模糊核作为输入值,运用NAS-RIF算法获取估计原始图像。根据图像匹配关系建立图像参考平面矩阵,设计虚拟现实图像融合算法。利用非局部均值原理算子,计算出图像中各像素的权重值,再结合图像稀疏度构建图像盲复原模型。仿真结果表明,所提算法复原后图像的平均结构相似度为0.94,结构相似度得到了提升,并且图像复原质量较高。If the weight of image pixels is ignored in the process of blind restoration of blurred images,the structural similarity of restored images will be low.Therefore,a blind restoration algorithm for blurred images in virtual reality scenes was proposed.First of all,the multi-scale method was adopted to decompose target images and estimate the fuzzy kernel at each layer of the image.After taking blur kernel as the input value,the NAS-RIF algorithm was used to estimate the original image.According to the matching relationship,the reference plane matrix of the image was constructed,and then the virtual reality fusion algorithm was designed.Moreover,the operator of the non-local mean principle was used to calculate the weight value of each pixel in images.Finally,the blind restoration model was constructed by combining with the image sparsity.Simulation results show that the average structural similarity of the images restored by the proposed algorithm is 0.94,which is improved.Meanwhile,the quality of image restoration is high.

关 键 词：虚拟现实技术 模糊图像 模糊核 盲复原 稀疏正则化 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]