Open Access proceedings Journal of Physics: Conference series

1.
 
Introduction 
Currently, research aimed at the development of methods for recognition, classification, and 
systematization of images of micro-objects, for example, pollen grains, medical unicellular organisms, 
images of useful minerals in the rock mass, fingerprints, is in great demand. The computer vision 
systems (CVS) of micro-objects will far exceed the capabilities of a person engaged in counting pollen 
in palynology, breeding and seed production, identifying reference points on the contour of images of 
useful minerals [1,2]. In the structure of the CVS, mechanisms of digital filtering, smoothing, 
recognition, and classification are used, based on the technology of obtaining and preparing 
microimages, input, encoding, compression, placing information in memory, selecting reference, and 
reduction of zero (redundant) points of the contour, representation and identification of images of 
micro-objects
[2]. In the structure of the CVS, mechanisms of digital filtering, smoothing, recognition, 
and classification are used, based on the technology of obtaining and preparing microimages, input, 
encoding, compression, placing information in memory, selecting reference, and reduction of zero 
(redundant) points of the contour, representation and identification of images of micro-objects. CVS, 
along with the performance of all these functions, organizes the exchange of data for microscopic 
studies, storage, which in traditional conditions are performed manually [3, 4]. This study is devoted 
to the development of software for the CVS, image database, and database, mechanisms for extracting 
statistical, dynamic, specific characteristics, extracting and dividing contours into segments, using 
templates for filtering, smoothing recognition, and classification of micro-objects. 

---

AMSD 2020
Journal of Physics: Conference Series