Evaluating Geometrical Parameters of Disperse ...

Dispersion is an essential part of many technologies, so automated evaluation of generated drops diameters and their distributions is one of the major problems at the stage of technological processes study and adjustment. The purpose of investigation is to develop a method of segmentation objects in the image, allowing improvement on particles recognition and efficient procession of their images in the presence of such contour defects as thin rays, fractal type shapes, presence a lot of objects in complex cases of their location in the image. The paper marks out the methods improving image quality, the use of which gives the best results for dispersed formations images considering their specificity. A modification of image anisotropic filtering method and algorithm to improve emulsions image quality have been designed, which can significantly improve the quality of high frequency noise eliminating and separating the drops from background elements. There is improved path tracing procedure for images containing a large number of objects in complex cases of their location. Method can work well with open curves. Efficient method for plane figures "fast" measurement has been proposed. Method is based on numerical integration formulas, taking into consideration the errors made by the computer as a measuring device. Statistical processing of drops size is made. The correspondence of empirical distribution to lognormal law is established. The software implementing the results of the investigation and combining all the stages of problem solving has been created. The results can be used to study liquids dispersing problems, making and using various dispersed formations, including new materials and equipment development.