Research On CCD Image Analysis Used In Optical Measurement Of Particle Parameters

With the increasing development of modern science and technology, spray systems are widely used in agricultural, industrial, medical and scientific applicants. Among all these fields, droplets produced by spray nozzles are the most common analytical objects. As the droplets size distribution reveals the transient behaviors and states of the spray field, the acquisition of droplets size distribution information is very important in spray device optimal atomization and industrial process control. As it is hard to calculate or predict the droplet size distribution due to the complexity of spray droplets forming mechanism, experimental measurements play an important role in this aspect. Among various corresponding measurement techniques, the droplets analysis method based on digital image processing has recently become a major research focus in the world.Firstly, a spray droplets measurement system is designed, including the spray system and droplets sampling and image detecting systems. Different spray droplets size distributions are obtained by the use of different kinds of nozzles under different conditions. Two kinds of image detecting systems are proposed, including the microscope image detecting system and the micro image detecting system. A novel image acquisition method is presented which has improved the existing methods. The droplets measurement system is combined with a rainbow measurement system, which gives calibration results to adjust the final results of the image processing method.Secondly, the pretreatment methods of transparent droplets images processing are analyzed. Several common threshold methods are compared. According to the features of the transparent droplets image, the noise and double edge problems are solved by small area elimination based on droplets contour marking. Several overlapping forms of the transparent droplets contours are studied, and the internal droplet filling method is used of which the shortage is overcome by a novel proposed contour gap closure method. This method has solved droplet contour inflection point phenomenon which is defined in this paper. An image background and particle removal method is proposed based on marking link, which gives the basis to the droplet contour rebuilding and size analysis.Thirdly, the vagueness phenomenon of the droplets image is analyzed, and some typical vagueness image blind restoration methods are compared. Several common image edge detecting methods are studied, and more accurate droplet edge is obtained by the use of a newly proposed edge detecting method which has combined inner boundary detecting and image thinning methods based on the analysis of morphology edge detecting. In order to solve the transparent droplet edge lacking problem, the contour of the droplet is fitted and rebuilt which restores the original shape of the droplet and gives size parameter.Lastly, the calibration method of both microscope and micro images is presented. The experimental results from image method and rainbow method are compared. a novel analysis method based on repeated droplet size statistics is proposed. Repeated droplets are recognized and removed according to repeating count reason, of which the validity and accuracy are proved by experiment. Finally, the measurement error is also analyzed and discussed.