Optimization of the processing of images of pollen grains based on the use of their specific characteristics and geometric features

Abstract

Methods have been developed to increase the accuracy of processing of images of pollen grains based on mechanisms for extracting statistical, dynamic, texture and specific characteristics, as well as geometric features of micro-objects. A technique is proposed to increase the accuracy of information processing by the metric characteristics of the contour points of the input and reference images. Methods have been developed for point and nonlinear verification of the correspondence of the contours of the input and reference objects based on the mechanisms of isolating, segmenting, interpolating, contrasting, extracting specific characteristics of pollen grains such as frequency, cytoplasm, reticular, spore, texture, morphology of the object and other geometric features of raster images. A set of information processing methods with mechanisms for reducing the zero points of the image contour, reducing the size of rasters, scaling, threshold and level control of dynamic parameters during image recognition and classification, adjusting the points of the color and brightness picture parameters, fixing the initial values, segment centroid and identification are proposed and implemented. Implemented a computer complex of micro objects view in C ++ in the parallel computing environment “CUDA”.