A Hybrid Solar Powered Chicken Disease Monitoring System using Decision Tree Models with Visual and Acoustic Imagery

Abstract

Avian diseases have been the prevalent cause of shortage in poultry demand. While preventive and control measures can be applied to address this concern, disease identification and verification has been a challenge – especially considering the fact that the modern solutions are costly and are not compatible for remote usage. With this, the researchers proposed the development of a hybrid solar-powered chicken disease monitoring system using decision tree models as the learning algorithm for identification using visual and acoustic inputs from the chickens. The learning algorithm identifies six different symptoms from chickens (slouching, eye foaming, lethargy, feather loss, color paling, and raling) from 15 different chicken samples. Throughout the experimentation of continuous monitoring for 72 hours, the learning model was calculated to have 84.6% accuracy in classifying diseased chickens when only visual imagery is considered, and 86.1% accuracy when audio inputs are also provided. With the integration of the hybrid solar-power system, the system can still continuously operate for 8 hour of power downtime.