A DMD-based Hyperspectral Imaging System Using Compressive Sensing Method

Sun Zhongqiu,Chen Bo,Cheng Chengqi
DOI: https://doi.org/10.1117/12.2068627
2014-01-01
Abstract:Hyperspectral Imaging Systems (HIS) are widely applied in many fields. However, in the traditional design of HIS, it is much time-consuming to acquire an integrated hyperspectral image. Compressive sensing is an efficient method to process sparse data, and a single-pixel camera which used the digital micromirror device (DMD) for accomplishing the CS algorithms had been developed. Nowadays, DMD achieved great development. The size of mirror array is increasing while switch speed of a single mirror becomes very fast. Consequently, researchers make efforts to design a HIS using CS method. CS method is a method to scale down the spatial information but the hyperspectral datacubes are still huge because of the thousands of bands. In this paper, we design a DMD-based spectrometer architecture using the method of compressed sensing principle, combined with DMD's spectral filter characteristics. In the new architecture, there are two DMDs. One is used for implementing the CS pattern, the other for filtering the bands. It has spectral simply adjustable advantages. With this new technology, we can reduce the amount of information which needs to be transmitted and processed in both spatial and spectral domain. We also present some simulation results of implementation procedures.
What problem does this paper attempt to address?