Investigation of the dispersion of mayonnaise products using a computer program

The paper proposes a new method for quantitative and qualitative analysis of the dispersion of mayonnaise emulsion using computer vision. The fat balls in the image are determined as an object and the area is calculated using the YOLO library (You Only Look Once) and a pre-trained user model. The user model was compiled on the basis of a pre-trained yolov8nseg model. Before ranking the fat balls, for the program to work correctly, it is necessary to take a photo from a microscope with a superimposed sample glass with a division price of at least 10 microns, which will allow you to obtain the actual size of the fat balls. Fat balls are ranked using interval series of variations and a histogram of frequencies already prescribed in the program. The developed program allows you to determine the quality of the dispersed phase of a mayonnaise emulsion in an express format, and with sufficient training of the user model, it will allow you to determine the dispersion of water-fat emulsions.