基于光场的分布式压缩感知  

Distributed compressed sensing based on light filed

作　　者：雷蕊[1] 沈为[1] 张之江[1] 牛涛[1] 

机构地区：[1]上海大学通信与信息工程学院,上海200072

出　　处：《电子测量技术》2014年第3期20-23,共4页Electronic Measurement Technology

摘　　要：针对光场渲染提出了一种分布式压缩感知重建算法——同时正则化的正交匹配追踪算法(simultaneous regularized orthogonal matching pursuit algorithm,SROMP)。光场相机阵列属于分布式信源,其获取的光场图像序列不仅具有图像内部相关性,还具有极强的图像间的相关性。验证了光场相机阵列信源特征符合分布式压缩感知第二联合稀疏模型。提出的SROMP算法,能利用信号间和信号内的相关性,高概率的同时重建多幅图像。并通过实验验证了算法的准确性。In this paper,we propose a distributed compressed sensing reconstruction algorithm based on the feature of light filed images-simultaneous regularized orthogonal matching pursuit algorithm （SROMP）.The light field camera array can be seen as a distributed source,the image sequence captured by the camera array contains the inter-correlation and intra-correlation.We test the light field images meet the second joint sparse model.SROMP can take advantage of the correlation to reconstruct light field image sequences simultaneously and high probably.Experimental results show that SROMP method is effective.

关 键 词：分布式压缩感知 光场 SROMP 

分 类 号：TN911.7[电子电信—通信与信息系统]