Work

A Survey of AI, IoT and Computer Vision Approaches for Livestock Health Monitoring and Management

Year: 2026
Type: preprint
Abstract: Monitoring the health of individual animals is difficult on farms where manual observation cannot be continuous, and late recognition of abnormal behaviour or illness raises treatment cost and lowers productivity. This paper surveys recent work on livestock health monitoring that combines Internet of Things (IoT) sensing, machine learning, deep-learning-based computer vision, and mobile or web applications. Representative studies on wearable biosensors, multi-sensor activity classification, federated learning, disease-oriented IoT systems, long-range wireless tracking, cloud and mobile monitoring, and image-based behaviour recognition are compared by sensing modality, analytics, user interface, and deployment constraints. The comparison shows that most systems address one capability at a time, such as vision-based behaviour recognition, sensor-based health prediction, or remote tracking, while integration with health, vaccination, feeding, and breeding records is seldom addressed. Guided by this gap, the survey outlines an integrated architecture in which computer vision, a temperature and humidity IoT node, a backend service, a record database, and mobile and web clients work together with expert validation of alerts, and it describes the measures needed to evaluate such a system. The architecture is being developed for a pilot at Krishi Vigyan Kendra (KVK), Baramati.
Language: English
Subfield: Plant Science
Open Access status: green
OpenAlex ID: W7236580147
Cites:
Cited by:
Related to: 10