Particle-sizing system for scanning electron microscope images of solid-propellant combustion exhaust

Abstract

Accurate measurement of particle size distribution of rocket motor exhaust is essential for predicting the combustion efficiency and infrared plume signature. This thesis presents an automated method for extracting particle size distribution from scanning electron microscope (SEM) images. The SEM images were taken off a filter paper placed the end of a collection probe inserted into the exhaust plume. The automated SEM extraction system consists of an IBM AT-based computer system fitted with a 512 x 480 pixel frame grabber. Photographic images taken off the SEM are acquired via a vidicon camera. A C language program was written to control the hardware and automate the extraction process. A threshold is first applied to the digitized image and the resulting binary image is subjected to object segmentation. Each object is then sized and the distribution from one or more images can be plotted. The main bulk of this thesis is to document the software specially written to undertake this set of tasks. Results obtained were compared with that from a Malvern MasterSizer particle sizer and found to be favorable. Particles as small as 1/8 um have been successfully sized.