Control preventivo de sigatoka negra en cultivo banano apoyado en redes convolucionales

his work focuses on solving the problem of classification and detection of black Sigatoka disease in banana plants, in terms of improving the process used in Colombia and reducing costs for the disease control process, based on the use of neural networks with the VGG19 Architecture. An automated tool is proposed for analysis, control and monitoring of Black Sigatoka. The analysis, control and monitoring will be done thanks to the reports as the final result of our tool, which will seek to be as explicit as possible for the end user in terms of location, severity and visualization of results classified in fields. The development of this project will base the use of tools for automation of processes in agriculture in Magdalena as it is based on real data and current deep learning techniques, exposing a vision of the use of precision agriculture as a set of techniques where the technology will begin to base decisions for the improvement of crops in terms of control, analysis and phytosanitary monitoring of diseases and associated fungi.

• Autores: Pallares, Carlos Jorge; Lallemand, Keneth Stive; Visbal, Fernando David
• Fecha de publicación: 2021
• Institución: Universidad del Norte
• Repositorio: Repositorio Uninorte
• Idioma: spa
• OAI Identifier: oai:manglar.uninorte.edu.co:10584/9533
• Acceso en línea: http://hdl.handle.net/10584/9533
• Palabra clave: Machine learning; Fourè; Segmentación; redes neuronales convolucionales; Sigatoka negra; enfermedad; foliar; preventivo; fitosanitario; agricultura de precisión; reducción de costos
• License: Universidad del Norte