Reconstruction of atmospheric vertical reflectivity profile on the base of discrete orthogonal polynomials

Because of their characteristic large contrast, the transfer problem of meteorological radar images is cumbersome and resource expensive. We present a simple and efficient process to code images based on our experience on discrete data fitting on circular and spherical domains. On either case, the disperse data may lead to large errors unless we use a discrete reduction. We consider, as is common in images analysis, a geometry independent single variable but discrete description, to reduce it using the Chebyshev like polynomials on that discrete set. We analyze several radar images of the type height-distance, to successfully deconstruct and reconstruct them with a reduced error.