Development of a computer vision algorithm using yolo convolutional neural networks for the classification of asparagus according to the bud

León León R.A.; Morales Salazar J.C.; Alcalde Portal M.B. · CISCI - Conferencia Iberoamericana en Sistemas, Cibernetica e Informatica (2024) · DOI: 10.54808/cisci2024.01.497

Resumen

The purpose of this research is to analyze the quality control inspection of the asparagus export process; This is done visually by experienced operators, however human vision does not follow a standardized selection criterion, so an artificial intelligence system was designed with convolutional neural networks based on YOLO v8 that detect the state of the tips or buds of the eyes. the asparagus according to its class, A which is closed tips, and class B which are open or flowered tips. Using an algorithm in Python v3.7 developed in Visual Studio Code, a dataset was created with 2,300 images of asparagus with the camera connected to the computer; of which 1610 images were entered for training, 460 for validation and 230 for testing the neural network, obtaining 96% accuracy according to the confusion matrix in the classification of the bud when testing in a controlled environment. © CISCI - Simposium Iberoamericano en Educacion, Cibernetica e Informatica 2024.

Instituciones: Universidad Privada del Norte