Cyber-Physical System for Determining the Presence of Adulteration in Meat Products Based on Automated Image Processing of Its Slices

A cyber-physical system for processing images of slices of meat products has been developed. The proposed processing procedure allows you to automate the procedure for detecting the presence of counterfeit. For this, an algorithm for solving the problems of forming images of counterfeit meat products and pattern recognition for detecting counterfeits is proposed. The mathematical model of images of slices of meat products is obtained on the basis of the transition from an RGB image to a variant of the Grayscale color model. To do this, it is proposed to use reference points, the coordinates of which are calculated using Gaussian images. The procedure is completed by determining the descriptors of reference points using the constructed histograms. The presence of falsification on the slice image is carried out by solving the recognition problem based on comparison with reference descriptors of reference points. When identifying signs of the presence of a falsification, the value of similarity with its standard is determined, which is calculated as the ratio of the number of points that match to the number of keypoints of the descriptors in percent. The characteristics of the information description, the functional structure of the proposed cyber-physical system, as well as the structure of user interfaces are given. In conclusion, the results of testing the system are considered, which confirm its performance.