Study to determine cloud motion from meteorological satellite data

Abstract

Processing techniques were tested for deducing cloud motion vectors from overlapped portions of pairs of pictures made from meteorological satellites. This was accomplished by programming and testing techniques for estimating pattern motion by means of cross correlation analysis with emphasis placed upon identifying and reducing errors resulting from various factors. Techniques were then selected and incorporated into a cloud motion determination program which included a routine which would select and prepare sample array pairs from the preprocessed test data. The program was then subjected to limited testing with data samples selected from the Nimbus 4 THIR data provided by the 11.5 micron channel.