基于频域的像素光场图像重聚焦算法  

Refocus of the pixel light field image in the frequency domain

作　　者：王宇[1] 张旭[1,2,3] 屠大维[1,2] 

机构地区：[1]上海大学机电工程与自动化学院,上海200072 [2]上海大学上海市智能制造及机器人重点实验室,上海200072 [3]上海大学机械系统与振动国家重点实验室,上海200240

出　　处：《光学仪器》2016年第2期121-127,共7页Optical Instruments

基　　金：国家自然科学基金资助项目(51205244)

摘　　要：提出一种基于像素光场的频域重聚焦算法。首先建立四维光场与像素光场的关系模型,然后基于傅里叶变换,建立了像素光场的重聚焦方法,并分析了傅里叶切片的作用和重采样的方法。在实验中,对比积分投影法,分析两种不同方法对同一场景的重聚焦效果,评价了两种方法的计算效率。实验结果表明,基于频域的傅里叶切片法在结果上等效于积分投影法,但是计算负担更小。This paper provides a refocus algorithm of the pixel light field in the frequency domain.It needs to pixelate the 4Dlight field data at first,then establishes the way of refocus in the Fourier transform and analyses the effect of Fourier slice and procedure of resample.Comparing with the method of integral projection in the experiment,the refocus results about the same scene of the two different ways are explained and the efficiency of each is evaluated.The experimental results show that the Fourier slice has a better computational performance which can achieve the same as the result with integral projection.

关 键 词：像素光场 重聚焦 傅里叶切片 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]