Fuzzy-logic Color Recognition System Using a Fast Defuzzifier

Abstract

Purpose of research. The research presented in this article is aimed at improving the accuracy of determining the color shade. The developed fuzzy-logical color recognition system was used as the subject of the study. The efficiency indicator was the result of calculating the sensitivity area percentage and RMSE of the developed method.Methods. A method based on fuzzy logic has been developed and implemented, namely, on the structure of Mamdani's fuzzy inference, which consists of the following stages: fuzzification, fuzzy logical inference, defuzzification. Triangular membership functions were used at the fuzzification stage. As a compositional rule, 12 input variables were used, combined on the basis of Zadeh's compositional rule in 27. At the defuzzification stage, the area ratio method was used. The object of the study was the developed mathematical model for determining color.Results. A mathematical model has been developed, consisting of 4 steps, which guarantees a clear definition of 9 colors and their shades. Based on the estimation of the root of the mean square error, it was concluded that the proposed model is better than traditional options. It is expressed by the fact that the developed method reacts on the interval of the entire surface of output variables, while traditional methods have dead zones to changes in input variables.Conclusion. A fuzzy-logical color recognition system was developed. In the course of experimental studies, it was found that the RMSE and sensitivity indicators have better results in relation to other systems.