A Study on Compressed Sensing of Ray-Space Data

在这项研究中，我们提出了一种新的圆形相机阵列光线空间压缩感知方法。压缩感知利用了射线空间的特殊特性，变换后的数据具有稀疏的数据结构。首先，我们定义了一种新的射线空间和新域之间的变换和反变换方法，该方法可以有效地减少信息的量。利用极坐标系下射线空间的特征，对其进行积分变换。通过引入逐次逼近，解决了计算量大的问题。在实验中，我们对变换后的数据在上述定义的域中进行了低频分量的约简，证实了原始射线空间数据可以以较小的误差重构。

In this research, we proposed a new compressive sensing method of Ray-Space in circular camera array. The compressive sensing utilizes special characteristics of Ray-Space that transformed data has sparse data structure. As the first step, we defined a new transform and inverse transform method between Ray-space and a new domain, where the amount of information can be efficiently reduced. We used an integral transform using the feature of Ray-space in polar coordinate system. We solved the problem of high computation cost by introducing successive approximation. In the experiment, we reduced low-frequency components of transformed data in the domain defined above, and confirmed that the original Ray-space data can be reconstructed with small error.