Image processing in automated measurements of raindrop size and distribution

A rapid method for evaluating raindrop size and size distribution has been developed. It is based on image processing with correlation analysis in the frequency domain. This technique has the advantage of being a direct measu rement method that automatically identifies and counts raindrops. Calibrati on was carried out using a standard image with known raindrop sizes. Drop s izes, ranging from less than 0.1 to over 85 mm in diameter, have been autom atically recognized and successfully measured. Error was not larger than 1. 5%. In addition, practical examples of use of the method for determining ch aracteristics of raindrops in rainfall are presented. (C) 1999 Elsevier Sci ence B.V. All rights reserved.