基于压缩感知的数字全息成像

Digital holography based on compressed sensing

对于具有高分辨力的数字全息来说,大量的数据冗余存在其中,这对后续的处理分析造成一定的负担。通过分析压缩感知技术,结合数字全息原理,利用压缩感知技术对全息图进行处理,在不损失全息图细节的情况下,减少数据量。为避免其余因素干扰,利用计算全息模拟全息图进行实验验证。对全息图进行多组压缩采样,并利用正交匹配追踪的贪婪算法进行重建。同时,利用傅里叶变换与卷积的重构方法对其进行光学全息重现,并通过归一化相关性对其进行比较。实验表明,压缩感知技术可以降低数字全息中存在的冗余,当采样50%时能较好地还原出全息图。

For high-resolution digital holography,the generated large amount of redundant data cause the burden for subsequent processing and analysis.Through analyzing the compressed sensing technology and combining the digital holographic principle,the hologram is processed without any loss of its details and the data amount is reduced.To avoid interference from other factors,the validation is conducted by using the computer generated holography to simulate the hologram.Various compressed sampling on holograms is performed,and the orthogonal matching pursuit greedy algorithm is used for reconstruction.Fourier transform and convolution reconstruction method is used to reproduce the hologram,and the results are compared with normalized correlation.The results show that compressed sensing technology can effectively reduce the digital holographic data redundancy,and the hologram can be better restored with 50% compressive sampling.