Determination of Droplet Size Distributions using Image Processing System

Abstract

This paper describes the image processing system that automatically analyzes the size distributions in droplet video images.
This processing system, based on a conventional video camera, a digital image processor, a 32 bit personal computer and a monitor, is operated by a software written by the Basic of program language.
Images are generated by taking the microphotographs of the water droplets captured on a plate of glass coated with a slightly deteriorated cylinder oil. After image acquisition, the recorded video frames are replayed and analyzed under computer control.
The performance data are given from the experiment which is carried out to compare measurements of real spray distribution obtained by using the automatic system with those manually obtained from a microphotograph of water droplets.
Consequently, the accuracy for the both measurements is 7.54 % in relative error for 290 water droplets used here being varied over the range of 3 to 136μm in diameter.
Furthermore, it is found that by utilizing the image processing algorithm, the optimum threshold level is selected and this level is the value just before the black shadow begins to appear in the background of a monitor.