Sistema de clasificación de tomates con visión artificial

Abstract

This research focuses on the development of a tomato classification machine using artificial vision techniques with the aim of improving the tomato classification process based on their quality. The study followed a methodology that involved several key steps. First, the specifications of the tomato classification system were detailed, following the NTC-1103-1 standard as a reference. Subsequently, a careful selection of the necessary hardware and software devices was conducted to ensure their suitability for image processing and analysis. The core of the research was the application of artificial vision techniques to carry out tomato classification. This phase focused on crucial criteria such as size, color, and homogeneity, which are essential visual attributes for determining tomato quality. The results revealed that the use of a neural network in this classification system provided better performance. The neural network demonstrated higher effectiveness in classifying high-quality tomatoes compared to traditional computer vision methods. This advancement represents a significant milestone in the automation of the agricultural product classification process and has the potential to reduce losses associated with manual classification, thereby improving the profitability of producers, and ensuring the delivery of high-quality products to consumers.