Compressive Spectral Imaging in View of Earth Observation Applications

Abstract: Earth observation from space is an important scientific and industrial activity that has applications in many sectors. The instruments employed are often large, complex, and expensive. In addition, they generate large amounts of data, which is challenging for storage and transfer purposes. Compressive spectral imaging would be a cheaper, more efficient, and well-adapted technique to perform Earth observation. An interesting architecture is compressive spectral imaging with diffractive lenses, which is extremely compact. This work investigates the possibility of replacing the diffractive lens in this system with a classical refractive lens. Taking advantage of the chromatic aberration of a lens makes the use of expensive diffractive lenses unnecessary. Simulations are performed to test the feasibility of the method. Signal recovery is a basis pursuit solved using the Douglas-Rashford algorithm.