像面编码计算成像实验研究及重构图像的质量评价

发布时间：2018-06-07 20:27

本文选题：计算成像 + 压缩感知　；参考：《红外与毫米波学报》2017年04期

【摘要】：针对国产化大面阵红外器件技术缺乏,用红外焦平面探测器获取图像时难以消除自身的非均匀性以及信噪比低,航空航天成像应用中的图像采集、传输、存储成本越来越高等问题.论文引入了像面编码计算成像,首先分析了这种成像系统的原理模型,然后将压缩感知理论应用于计算成像中,搭建成像原理样机,进行非压缩和压缩的计算成像实验.最后在重构的图像质量评价中引入了信号子空间分析方法,进行重构图像的信噪比估计,实验结果表明,这种信噪比估计方法更加准确有效,并且可以以此作为实验结论给出实际压缩成像时需要的合理采样次数.
[Abstract]:In view of the lack of domestic large area array infrared device technology, it is difficult to eliminate the non-uniformity and low signal-to-noise ratio when using infrared focal plane detector to obtain image, so the image acquisition and transmission in aerospace imaging application. Storage costs are increasing and so on. Firstly, the principle model of the imaging system is analyzed, then the compression sensing theory is applied to the computational imaging, and the prototype of the imaging principle is built, and the non-compression and compression computational imaging experiments are carried out. Finally, the signal subspace analysis method is introduced into the reconstructed image quality evaluation to estimate the SNR of reconstructed image. The experimental results show that the SNR estimation method is more accurate and effective. And it can be used as the experimental conclusion to give the reasonable sampling times needed in the actual compression imaging.
【作者单位】：中国科学院上海技术物理研究所中科院空间主动光电技术重点实验室;中国科学院大学;
【基金】：上海技术物理研究所创新专项(DX38)~~
【分类号】：TP391.41

【相似文献】