A simple technique for sizing and counting spray drops using digital image processing

Abstract An efficient and economic system for measuring the size of spherical spray drops was developed and tested. A simple patter recognition technique peculiar to circles was used in combination with a conventional immersion sampling and photographic method. The system was designed to size and count a relatively large number of drops in a short time, and also to filter out images of undesirable objects such as drops in contact of overlapping each other, as well as odd-shaped foreign materials. Proper choices of the magnification in constructing an image frame and of a threshold level to convert the original image to a binary digital image were important. The image processing system described considerably reduces the total measuring time, eliminates subjective observer error in sizing and counting spray drops, and has proved to be substantially efficient.