Desarrollo de un Algoritmo de Visión Artificial para la Clasificación de Fresas de Exportación o Consumo Nacional

This article developed an algorithm using computer vision employing a convolutional neural network with YOLO to classify strawberries for export and domestic consumption. This is crucial as export companies strive daily for the proper collection, sorting, and disposition of strawberries to enhance profitability. The tests were conducted on a Lenovo laptop with an Intel i7 processor and Windows 11. A Logitech c920 camera was used to detect the strawberry's coloration, which was integrated into the programming done in Visual Studio Code. The YOLOv5 network, specifically the YOLOv5x model, was employed, pre-trained with images collected by the research team. The training was done in Google Colab before integrating the neural network into the programming. After conducting various tests, an overall efficiency of 97.14% was achieved for both classifications with a margin of error of 2.86%. This outperforms other works, such as the thesis "Classification of apples using computer vision and neural networks," which attained an efficiency of 92.25%. Our results suggest that the application of this image processing system with neural networks will simplify processes and gradually increase productivity within the production line.