RECOLECCIÓN Y RECONOCIMIENTO DE EVENTOS DOMÉSTICOS MEDIANTE SENSORES DE SONIDO EN IOT
Fecha
2021-09-09
Autores
Título de la revista
ISSN de la revista
Título del volumen
Editor
Jaén: Universidad de Jaén
Resumen
El objetivo de este TFG es el desarrollo de un disposito IoT que permita reconocer eventos de sonido en un entorno
doméstico. Para ello, en primer lugar, se realizará una recolección de muestras de sonido sobre diferentes eventos
diarios que puedan ser recogidos el dispositivo de sonido, tales como apertura de grifos de agua, puertas,
electrodomésticos, etc... En segundo lugar se procederá a construir un modelo basado en Deep Learning que permita
clasificar los sonidos y sea compatible con una ejecución en el dispositivo IoT, obteniendo una primera evaluación de
su capacidad. En último lugar, se evaluará la posibilidad de integrar el modelo en tiempo real incluyendo un caso de
estudio de una escena real.
The objective of this TFG is the development of an IoT device that allows to recognize sound events in a domestic environment. To do this, in the first place, a collection of sound samples will be carried out on different daily events that can be collected by the sound device, such as opening of water taps, doors, electrical appliances, etc ... Secondly, we will proceed to build a model based on Deep Learning that allows to classify sounds and is compatible with an execution in the IoT device, obtaining a first evaluation of its capacity. Lastly, the possibility of integrating the model in real time will be evaluated including a case study of a real scene.
The objective of this TFG is the development of an IoT device that allows to recognize sound events in a domestic environment. To do this, in the first place, a collection of sound samples will be carried out on different daily events that can be collected by the sound device, such as opening of water taps, doors, electrical appliances, etc ... Secondly, we will proceed to build a model based on Deep Learning that allows to classify sounds and is compatible with an execution in the IoT device, obtaining a first evaluation of its capacity. Lastly, the possibility of integrating the model in real time will be evaluated including a case study of a real scene.
Descripción
Palabras clave
Tecnología de la información / Tratamiento inteligente de información