连续推扫计算光谱成像技术

Continuous Pushbroom Computational Imaging Spectrometry

计算光谱成像技术具有高通量、快照成像等优点,但快照成像采样数据量不足,导致利用压缩感知方法重构图谱精度很低。通过对计算光谱成像技术各个环节进行系统研究,提出一种新型的连续推扫计算光谱成像技术,利用正交循环编码孔径代替传统的随机编码孔径,通过逐行扫描方式及正交变换可完整重构图谱数据。仿真和实际成像结果表明,连续推扫计算光谱成像技术可消除图谱混叠影响,理论上可完全重构图谱信息,重构图谱精度明显优于传统的计算光谱成像技术。相比国际上提出的多次曝光计算光谱成像技术,连续推扫计算光谱成像技术不需要改变编码孔径与探测器间的相对位置,也不需要凝视成像,系统中没有活动元件,稳定性高,适用于常规航空航天遥感推扫成像。

Computational imaging spectrometry(CIS)has drawn great attention in recent years.It has the advantages of high optical throughput,snapshot imaging and so on.On the other hand,CIS has the disadvantage ofinsufficiency in sparse sampling which reduce the accuracy of reconstructed spatial-spectral data.By analyzing the optical property of CIS,a continuous pushbroom computational imaging spectrometry(CPCIS)is presented.In CPCIS,the orthogonal cyclic coded aperture was used,and the continuous scanning line by linewas implemented through platform moving.The entire spatial-spectral data wasreconstructed by orthogonal inversion.According to the imaging simulation and experiment,the aliasing in spatial-spectral image was eliminated,and the reconstructed image was well satisfied.Comparing to multiframe CIS,CPCIS has no moveable element,which can image without staring the object,thus it is suitable for the airborne and spaceborne remote sensing applications.