Abstract

The efficacy of a new algorithm implementing the minimum-negativity constraint¹ in computing data sets of up to 64K (K = 1024) data points is demonstrated. A specific example is shown for an interferogram of 17K points taken from a rocketborne field-widened interferometer. Restoring data points to the interferogram to yield a total of 64K points, on Fourier transforming with the fast Fourier transform, produces improved resolution for the resulting 64K data points of the IR spectral lines. An example of restoration with the same interferogram involving a minor modification of the algorithm so that it is appropriate for parallel processing is also illustrated. The new algorithms may be easily generalized to two dimensions.

© 1987 Optical Society of America