Blind Deconvolution of Long Exposure Lens-Based Chromotomographic Spectrometer Data

摘要

Projection-based chromotomographic spectrometers are sensors that collect both spatial and spectral information with fairly simple optical as well as electronic hardware. Efforts to utilize them for remote sensing applications have met with obstacles primarily due to the fact that the impulse response of the imaging system as a function of wavelength must be know in order to reconstruct the spatial/spectral content of the scene under study. This paper reports a blind deconvolution algorithm specifically designed to reconstruct the spectrum of the scene under study as well as an estimate of the wavelength dependent atmospheric transfer function of the system. The method is tested using simulated data with realistic turbulence and noise factors in order to demonstrate its effectiveness.