Diseño e implementación de un controlador basado en reconocimiento de patrones para mejorar el abastecimiento de insumos alimentarios en galpones avícolas.

Abstract

The aim of this research was to design and implement a controller based on pattern recognition, to improve the supply of food inputs in poultry sheds. The research is quasi-experimental because the tests environment is previously defined, for this reason the research is descriptive and applicative, for the collection of information a controller was implemented for a section of 10 chickens. Such controller uses an artificial vision methodology for decision making applying fuzzy logic, which interprets whether or not there is a presence of chickens, the quantity and type of food, according to the size and number of birds using the OpenCV classifier. In the comparison of hypothesis using Chi-square test was obtained a value of 1.35 which indicates that the differences are not highly significant, however, the benefits in relation to the time and human presence in the shed are notable. Considering that one of the most fundamental points in the current research is the recognition of patterns the same was assessed both in an unsectioned environment with a precision of 38.9% and in a sectioned environment which according to the manual AGROCALIDAD version 2016, with a recognition accuracy of 91.25%, that is why it is recommended to use cameras, software and hardware of the best features for optimal recognition considering that of this parameter the food supplies provisioning in the poultry shed depends on.