Sensitivity analysis of parameters affecting image processing of high pressure gaseous jet images

Digital image processing is one of efficient methods that is adopted today for the determination of fuel spray characteristics in internal combustion engines. The imaging method and the images quality are two important factors that influence the accuracy of results. In this paper, an edge finding method was employed for processing of the Schlieren image of the gaseous jet, and sensitivity analysis has been conducted to investigate effects of the edge finding method threshold, the experiments scale factor and the level of the blur and the noise of images on image processing results. A method based on the Otsu method has been proposed also for finding the dynamic threshold for Schlieren jet images. Results showed that the sensitivity of the jet tip velocity to the edge finding method threshold was low (in the valid range of the threshold), but in regions that the jet tip velocity was low, this sensitivity was as high as 14%. The sensitivity of the jet tip penetration to the blur level of images was less than 2%, while the tip velocity was more sensitive to the blur content of images, and the blur with the standard deviation of 4 could led to 28% error in the tip velocity.