DEVELOPMENT QUALITY CONTROL SYSTEM THROUGH ARTIFICIAL VISION FOR THE DETECTION OF STAINS GENERATED BY CELL BREAKS IN CANNED ARTICHOKE (#793)

Authors

León León, Ryan Abraham

Alvarado Avalos, Irvin Eduardo

Abstract

The present research work put into practice the new technologies related to artificial vision so that, through the use of neural networks, specifically convolutional ones, the general objective can be achieved by developing a quality control system through artificial vision to be able to detect the different stains generated by cellular breaks that may exist in preserved artichokes. In order to execute artificial vision, the project was supported by the Python program, which made it easier to handle the programming language together with PyCharm, both of which are an ideal complement so that correct coding and elaboration of the desired system can be developed. . Furthermore, artificial vision is considered one of the most efficient methods, since, in its development, it works with a large visual database, which, in the training stage, is allowed to learn and subsequently be able to predict and recognize on its own what you are showing it, generating the desired detection. However, this use is linked to constant research so that it can work correctly. Finally, after making a precision table, it showed that the system manages with 98% of this, compared to other investigations that manage percentages of 83.9% and 96.6%. This allows us to conclude that the work carried out based on a large collection of data works correctly, allowing its future implementation in the quality area.