Published Paper Details:

Livestock farming plays a vital role in agricultural productivity, food security, and rural economic development, but continuous animal health monitoring remains difficult because traditional supervision depends on manual observation and delayed veterinary response. This study presents an intelligent livestock health and activity monitoring system using deep learning, embedded automation, and IoT communication for early disease detection and farm management. Images of healthy cattle and cattle affected by Lumpy Skin Disease are collected from a Kaggle dataset and used to train a CNN-VGG19 classification model. The trained model is deployed in a Flask based web application for real time disease prediction using uploaded animal images. Prediction results are transmitted to an Arduino UNO through serial communication for automated decision making. When disease is detected, the system activates a chemical sprayer and displays alerts on an LCD module. Healthy cattle continue under normal monitoring conditions. Additional features include RFID based boundary monitoring, manual food and water control through the Blynk application, and LCD based system status display. Experimental analysis demonstrates improved disease response efficiency, reduced labor dependency, and effective livestock safety monitoring for modern animal husbandry applications. The proposed framework supports reliable automated livestock supervision in small and medium scale farms efficiently.