Eventos Anais de eventos
ENCIT 2022
19th Brazilian Congress of Thermal Sciences and Engineering
An Image-Based Technique for Ice Quality Evaluation
Submission Author:
Fabio de Souza Melo , SC
Co-Authors:
Fabio de Souza Melo, Vitor Alexandre Massochin, Christian Hermes, Rodrigo Perito Cardoso
Presenter: Fabio de Souza Melo
doi://10.26678/ABCM.ENCIT2022.CIT22-0252
Abstract
Due to the upcoming demand for ice that lasts longer and has a better appearance, clear ice has become a target for the refrigerator manufacturers. Although qualitative evaluations have been adopted over the years, particularly on industrial grounds, development of cutting-edge techniques for high-quality ice production relies on quantitative measurements, which are not only more accurate but also less subjective than the qualitative ones. Therefore, in this study, an image-based ice quality evaluation technique is proposed. A purpose-built apparatus was especially designed and constructed in such a way that a controlled environment and a professional camera are used to take pictures from the faces of a 125-cc ice cube. After this step, the images undergo a thresholding procedure, where the RGB matrix of the picture is transformed into grayscale and then binarized. Such an approach was able to identify the regions with flaws and, therefore, to calculate a figure of merit representative of the ice clearness, which is expressed by grades spanning from 0 (opaque) to 10 (translucent). Such a methodology was validated using ice cubes with distinct quality produced by means of another purpose-built apparatus that controls the freezing rate as well as the water agitation during ice making.
Keywords
Ice quality evaluation, Clear ice, Ice making, Image-based technique
