Information Features for the Computer Classification of Pollen Grains

Abstract

In this paper, information features for the computer classification of pollen grains are considered, and their morphological structure is analyzed with consideration of geometric features at different levels. The dimensions (equatorial diameter and polar radius, etc.), outlines, sculpture, aperture, texture, polarity of the pollen grains were revealed as recognition descriptors. The example of pollen grain classification is given. The intervals and the number of plants that fall within these intervals according to the characteristics (the equatorial diameter and the polar radius), were found. For the equatorial diameter, it was found that the maximum number of plants (25 pieces) lies in the range from 40 to 50 μm, etc. According to the findings, a table of ranges of the equatorial diameter and the number of plants falling within the data range and a diagram were constructed.It was found that no plants fell within the range of values of the polar radius from 0 to 10 μm, 6 plants fell into the range from 10 to 20 μm, from 28 to 20 μm to 28 μm, etc. Based on the data, a table of the ranges of the polar radius and the number of plants falling within the data range and a diagram were constructed.