Analysis Of Engine Spray Image Filtering And Processing Algorithms

Fuel atomization has a decisive role in the engine of the economy, power, emissions, so the research of spray process has important significance. Flash photography is a commonly used method of research on fuel atomization quality, but this method need to pick out a few good shooting images from a large image of the spray. If for each operating condition through artificial selection, time-consuming and laborious, without a certain standard, there are artificial judgment errors, so it is necessary to design a set of automatic spray image screening system to realize the automatic spray image screening.Aiming at the problem of preliminary screening in the spray images about over exposure, under exposure, uneven brightness, the image blur, select gray histogram, the average gray value, the local average gray level, the gray mean grads operator, Laplace and so on the different methods to filter out undesirable sequence image.In the preprocessing stage, through the comparison of the median filter and Butterworth low-pass filter on the spray image processing effects, we finally choose median filtering. In the particle recognition stage, mainly compares the advantages and disadvantages of Harris algorithm and Graythresh algorithm, we finally choose Graythresh algorithm. Determine the particle uniformity through the use of STD function to get the number of particles in different regions of the spray images of standard deviation, and so we measure the uniformity of the particle distribution.Finally according to the preset algorithm, the spray image screening system is designed in the MATLAB/GUI, and we will add it to the spray particle image measurement system to make it more perfect. We process the Spray images which are selected by the screening system and compare the results with Malvern, in order to test the accuracy of the screening system for screening results.