Examinando por Autor "Leal Lara, Daniel David"
Mostrando 1 - 2 de 2
- Resultados por página
- Opciones de ordenación
Ítem Modelo para detección de Xanthomonas campestris aplicando técnicas de machine learning mejoradas mediante algoritmos de optimización.Leal Lara, Daniel David; Barón Velandia, JulioThis proposal focuses on the elaboration of a model that allows the early detection of the Xanthomonas Campestris disease by applying Machine Learning techniques, characterized by their high interpretability, improved by means of optimization algorithms, allowing to accurately identify the state of a plant (Healthy or diseased), so that farmers can take early action reducing the impact generated by the disease in the presentation and yield of the crop.Ítem Modelo para la detección de la enfermedad Xanthomonas campestris en hojas de judía aplicando algoritmos de optimización genéticos y de gradienteRocha Calderón, Camilo Enrique; Leal Lara, Daniel David; Barón Velandia, JulioThis document presents the results of the elaboration of 8 models based on fuzzy logic systems improved by optimization algorithms, in order to provide an alternative solution to the early detection of Xanthomonas Campestris disease found in the leaves of the plants of bean (Kidney beans), allowing to properly identify the state of a plant (healthy or diseased). The models are obtained from the creation of 2 systems of fuzzy logic type Mamdani and Sugeno with different configurations in their input sets and in their rules, each of the configurations is improved using optimization algorithms, which use Exact methods (Quasi-Newton) and heuristics (Genetic algorithms). The methodological technique applied for the implementation of the models is based on the set of data or images to be analyzed and on the most relevant variables included in the color intensity of the RGB scale, by means of which the sets are defined of adequate classification about the state of the plant. The result obtained for the best model shows a performance of 99.68%, through its evaluation with the training data set, on the other hand, it provided a percentage of effectiveness of 94% in detecting Xanthomonas Campestris disease in a Bean leaf represented by an image, based on the test data set, allowing earlier detection of the disease in relation to conventional methods, so that farmers can take actions to reduce the impact of the disease in the presentation and yield of the crop.