Automatic Cloud Classification
Babu M. Mehtre, N N Murthy, Bojan Lipovscak, B. Chatterjee · IETE Technical Review · 1986
In this paper, design and implementation of an automatic cloud classifier is discussed. It is assumed that the pattern classes obey the multivariate normal distribution. Maximum Likelihood with threshold and penalised Declassification algorithms have been implemented both on host computer VAX-11/750 and Array Processor FPS-100. This technique classifies not only different cloud types but land, sea and snow as well. The result of classification and the time required on host as well as on array processor for various complexities is presented. NOAA satellite pictures have been used for testing.