Prototipo de identificación del mosquito AEDES con TINYML
No Thumbnail Available
Date
2025
Journal Title
Journal ISSN
Volume Title
Publisher
PUCE - Ibarra
Abstract
Este proyecto presenta el desarrollo de un sistema innovador para la identificación automática del mosquito Aedes mediante la implementación de tecnología TinyML. El sistema integra hardware especializado para la captura de señales acústicas emitidas por el mosquito, un modelo de aprendizaje automático para dispositivos de bajo consumo, además se desarrolla una aplicación móvil para la visualización y gestión de datos. La solución propuesta utiliza la plataforma Edge Impulse para el entrenamiento y optimización del modelo, permitiendo una identificación precisa y en tiempo real del mosquito. El prototipo está diseñado para operar de manera autónoma en condiciones de campo, contribuyendo significativamente a la vigilancia entomológica y el control vectorial. Este prototipo busca optimizar las estrategias actuales de prevención y control de enfermedades transmitidas por el Aedes, como el dengue, zika y chikungunya, especialmente en regiones con recursos limitados. La detección temprana de mosquitos permite a las autoridades sanitarias tomar decisiones informadas de manera rápida y efectiva, como la implementación de medidas de control focalizadas o la alerta a la población en zonas de riesgo. Se espera que este prototipo contribuya a reducir la incidencia de enfermedades transmitidas por el Aedes y a mejorar la calidad de vida de las comunidades, especialmente en aquellas con recursos limitados.
This project presents the development of an innovative system for the automatic identification of the Aedes mosquito by implementing TinyML technology. The system integrates specialized hardware to capture the mosquito's acoustic signals, a machine learning model for low-power devices, and a mobile application for data visualization and management. The proposed solution uses the Edge Impulse platform for model training and optimization, enabling accurate and real-time identification of the mosquito. The prototype is designed to operate autonomously in field conditions, significantly contributing to entomological surveillance and vector control. This prototype seeks to optimize current strategies for the prevention and control of Aedes-borne diseases, such as dengue, zika, and chikungunya, especially in resource limited regions. Early detection of mosquitoes allows health authorities to make informed decisions quickly and effectively, such as implementing targeted control measures or alerting the population in at-risk areas. It is expected that this prototype will contribute to reducing the incidence of Aedes-borne diseases and improving the quality of life of communities, especially in those with limited resources.
This project presents the development of an innovative system for the automatic identification of the Aedes mosquito by implementing TinyML technology. The system integrates specialized hardware to capture the mosquito's acoustic signals, a machine learning model for low-power devices, and a mobile application for data visualization and management. The proposed solution uses the Edge Impulse platform for model training and optimization, enabling accurate and real-time identification of the mosquito. The prototype is designed to operate autonomously in field conditions, significantly contributing to entomological surveillance and vector control. This prototype seeks to optimize current strategies for the prevention and control of Aedes-borne diseases, such as dengue, zika, and chikungunya, especially in resource limited regions. Early detection of mosquitoes allows health authorities to make informed decisions quickly and effectively, such as implementing targeted control measures or alerting the population in at-risk areas. It is expected that this prototype will contribute to reducing the incidence of Aedes-borne diseases and improving the quality of life of communities, especially in those with limited resources.
Description
Keywords
Aedes, Tinyml, Edge impulse, Identificación acústica, Vigilancia entomológica, Aprendizaje automático, Salud pública, Aplicación móvil