Assessment of the Influence of Storage Conditions and Time on Red Currants (Ribes rubrum L.) Using Image Processing and Traditional Machine Learning

Abstract

This study was aimed at revealing the usefulness of the combination of image analysis and artificial intelligence in assessing the quality of red currants in terms of external structure changes under the influence of different storage conditions. Red currants after harvest were subjected to storage at room temperature and at a lower temperature in the refrigerator for one week and two weeks. The statistically significant differences in selected image textures as a result of prolonged storage were determined for both samples stored in the room and the refrigerator. However, the changes in the structure of the red currant samples stored at room temperature were greater than for storage in the refrigerator. Distinguishing samples using models built using machine learning algorithms confirmed the usefulness of selected textures to assess the influence of storage conditions and time on red currants. Unstored red currants, samples stored at room temperature for one week, and those stored at room temperature for two weeks were classified with an accuracy of 99–100%, and unstored samples, fruit stored in the refrigerator for one week, and that stored in the refrigerator for two weeks were correctly distinguished at an accuracy of 97–100%, depending on the algorithm. Models developed for distinguishing red currants stored at room temperature and in the refrigerator for one week provided an accuracy of 99–100%, and for the classification of red currants stored at room temperature and in the refrigerator for two weeks, an accuracy equal to 100% for all used algorithms was determined.