Research On Computational Imaging Spectroscopy

Recently, a novel imaging spectrometry called computational imaging spectrometry has been proposed. Based on the conventional dispersive imaging spectrometry, a coded mask which modulates and compresses the3D spatial-spectral data-cube about the scene is imported appropriately in the light path. The3D spatial-spectral information about a scene of interest is first encoded and captured with one snapshot at the two-dimensional (2D) detector array. Compressed sensing (CS) theory is then used to reconstruct the3D data-cube from the2D aliasing image. Compared to the conventional imaging spectrometry, computational imaging spectrometry eliminates the disadvantages of low light throughput and temporal scanning, greatly reduces the amount of original data, and alleviates the pressure of the data storage and transmission. Systematic study of computational imaging spectrometry is carried out in this paper, and the main work is included as follows:Firstly, the classification, applications and bottleneck of conventional imaging spectrometers are summarized. The development of computational imaging spectrometry is described.Secondly, the basic theory of compressive sensing and computational imaging spectrometry is introduced and the mathematic model of computational imaging spectrometry is established. The key technology of data inversion is analyzed comparatively.Thirdly, the selection of coded mask and dispersive element is introduced. Several kinds of errors are simulated in the imaging chain of computational imaging spectrometry, and the effect of smile and key-stone aberration of the prism on the system is emphatically analyzed.Fourthly, imaging quality evaluation method of computational imaging spectrometry is proposed. Combined with the principle of computational imaging spectrometry and compressive sensing theory, the slated-edge method is used to calculate the modulation transfer function (MTF) of the reconstructed chart image and the distortion of the recovered spectrum is considered. Then evaluation factor is defined, and the aliasing spectral number, which affects imaging quality of the system, is quantitatively analyzed.Fifthly, the experimental device is designed and fabricated. The calibration of the system and imaging experiment are carried out. Through spatial-spectral information reconstruction, the capture of data-cube from a snapshot is achieved.Lastly, a so-called piecewise continuous coding computational imaging spectrometry is described. The first image plane of computational imaging spectrometer is segmented into several continuous spectral sub-band areas. The sparse sampling of the entire spatial-spectral image could be acquired by scanning. The simulation shows the reconstruction accuracy is significantly improved.