DEVELOPMENT OF A STATISTICAL MODEL FOR THE SYSTEM OF INSTRUMENTAL SORTING OF MINK SKINS BY COLOUR-DIFFERENCE PARAMETERS

Abstract

The creation of a system of instrumental quantitative assessment of the hair colour of mink skins is a topical and timely task, due to the fact that the methods of visual assessment of the colour of skins during sorting do not meet the current level of scientific and technological progress. The authors used statistical analysis methods to show the possibility of colour separation of mink skins using the colour parameters of the CIE L*a*b*colourimetric system. It is found that the greatest contribution to the overall discrimination, and hence to the colour difference, is made by L*, lightness; the variable b* is second in this process, and the variable a* is characterised by the smallest contribution. The developed classification matrix allows us to summarise the information about the number and percentage of correctly classified cases for each colour type and assess the degree of compliance of the classification results obtained by formal discrimination with the actual distribution of skins. A statistical model of the system of instrumental sorting of mink skins by colour-difference parameters is proposed, which allows evaluating the correctness of assigning mink skins selected by visual expert method to certain colour types.