Sistema de inspección y clasificación de hojas de plantas medicinales por medio de visión artificial

This text exposes the development of a system for the classification of leaves of medicinal plants, based on artificial vision and with the purpose of being used in a plant of extraction of oils and elaboration of tisane. Initially the process has two options, the first can determine what kind of plant is within the system (calendula, fennel), making the process automatically, the seconds options is the possibility of the operator selecting the type of plant that is is currently in the process; The current state of the plant is categorized into three processes (process_1: oil extraction, process_2: tisane, process_3: waste), to decide which is the most appropriate process to implement different classification methods, such as segmentation algorithms by means of of k -media, MLP neural networks and convolutional neural networks, these methods are programmed in python and mounted in an embedded system (Raspberry Pi 3), for their ease of use and for low cost, compared to other embedded systems that although they are more complete , they have higher costs. The process begins with a conveyor belt that carries the plants (calendula and hinojo) independently, after this they arrive at the inspection stage, where a photograph is taken and the image is processed, finally depending on the quality evidenced by the system. It is decided whether it is used for oil extraction or for tisane. The objective of this automation is to minimize times and increase the quality of the products, since the technification of this process is currently incipient; It is important to consider that at this moment the work is carried out manually, where the selection phase is carried out by a person, which leaves the decision of the chosen plant state to the subjectivity of each individual.

• Autores: Gaviria Montoya, Rogger David; Marin Hurtado, Cristhian David
• Tipo de recurso: Trabajo de grado de pregrado
• Fecha de publicación: 2018
• Institución: Universidad Autónoma de Occidente
• Repositorio: RED: Repositorio Educativo Digital UAO
• Idioma: spa
• OAI Identifier: oai:red.uao.edu.co:10614/10529
• Acceso en línea: http://hdl.handle.net/10614/10529
• Palabra clave: Ingenieria Mecatrónica; Sistema de inspección y clasificación de hojas de plantas medicinales por medio de visión artificial; Procesamiento de imágenes; Visión artificial; Algoritmos; Python
• Rights: openAccess
• License: Derechos Reservados - Universidad Autónoma de Occidente