Compressive Sensing Multi-spectral 3D Imaging

Abstract

In order to realize high efficient compressive sensing(CS) three-dimensional(3D) multi-spectral imaging, a 3D multi-spectral imaging system based on CS theory as well as the corresponding CS algorithm to reconstruct the 2D images based on joint optimization and the 3D construction algorithm based on photometric stereo is proposed in this paper. In the system a digital micromirror device(DMD) is used as spatial light moderator and a liquid crystal tunable filter(LCTF) is used to choose specific wavelength to pass. The object is measured with random patterns from different directions and the 2D images are reconstructed by the algorithm with the joint constraints, and the 3D model of the object is constructed from the 2D images via photometric stereo method. The multi-spectral and 3D images are accurately reconstructed at 0.3 mean sampling rate and the precision of both spatial and spectral is higher than the algorithm without joint constraint method using the same measurements.