Prototype for the detection and classification of food products using artificial vision based on color

Abstract

This work showcases a prototype that uses artificial vision to enhance manufacturing processes that require the classification of products. This can include detecting defects, ensuring compliance with standards, and categorizing products based on characteristics such as weight, contour, and color. The project aims to develop a prototype for the detection and classification of food products using artificial vision, where color serves as a reference point. To develop the prototype, the components for the design are chosen, and the characteristics of the object to be controlled are identified. In this case, the chosen characteristic is color. An algorithm is then created to enable the prototype to detect  the object’s feature. For this purpose, the programming language Python is used, which is governed by a free software license and has a library called OpenCV. OpenCV is an open-source computer vision library that contains around 2,500 algorithms and 500 functions. In the next stage, the prototype components, including a web camera, lighting source, computer or mini controller, image processing software, and display screen, are integrated. Real-time tests are conducted to validate the prototype, resulting in a functional prototype that meets the stated objective. This proves that artificial vision-based quality control systems can be adapted to manufacturing processes, leading to improved productivity and competitiveness.