SEDL / STP / STP848-EB / STP32621S

An Instrumentation System to Automate the Analysis of Fuel-Spray Images Using Computer Vision

Oberdier, LM
Staff research engineer, General Motors Research Laboratories, Warren, MI

Pages: 14    Published: Jan 1984

Download this paper for $25 PDF (284K)          View License Agreement
Abstract

An intelligent instrumentation system is being implemented to analyze images stored during pulsed laser illumination of fuel sprays. Images are stored on magnetic video disk during the experiment and later analyzed with a system consisting of a commercial image processor and a minicomputer. The system employs computer vision algorithms to reduce the effects of shading and to segment candidate droplets from each image. Pattern recognition methods will be used to extract features from each droplet and to reject those which are not in the sample volume. The final output of the system is an accumulation of droplet size distributions at various sample points in the spray.