基于压缩感知的光场获取与重建  

Reconstruction and capturing of light field based on compressive sensing

作　　者：张金远[1] 王红梅[1] 岳贤军[1] 张之江[1] 

机构地区：[1]上海大学特种光纤和光学获取重点实验室,上海200072

出　　处：《电子测量技术》2013年第2期51-54,共4页Electronic Measurement Technology

摘　　要：光场图像经常被用于产生三维场景,尽可能多的光场图像能够更精确的还原出真实的三维场景,然而这样就会使数据量过大,图片获取成本更高。提出一种新的方法,通过特定的光场图像获取方式,结合贝叶斯重建算法,以及光场渲染技术,不仅可以还原出原始视点的光场图像,更可以渲染出新的视点。此外通过运用图像压缩感知算法,使得该方法能够减少实验数据量、提高运算速度。实验结果表明该方法能够还原出高信噪比的原始视点图像——误差小于0.3dB,以及新视点图像。Light field images are used to reconstruct 3D images.The more images are used,the better results are recontructed.However,this leads to huge amounts of data and high cost of the images acquisition.This paper proposes a new method to reconstruct orignal light field images and renders new light field images based on compression sensing with a special way of capturing light field images and Bayesian reconstruction.The amount of data can be reduced and the efficiency can be improved.The experiments has proved that the error of rendering new light field images is less than 0.3 dB.

关 键 词：光场渲染 压缩感知 贝叶斯重建 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]