Airborne Vision-Based Remote Sensing Imagery Datasets From Large Farms Using Autonomous Drones For Monitoring Livestock

Abstract

Livestock have high economic value and monitoring of them in large farms regularly is a labour-intensive task and costly. The emergence of smart data on individual animals and their surroundings opens up new opportunities for early detection and disease prevention, better animal care and traceability, better sustainability and farm economics. Precision Livestock Farming (PLF) relies on the constant and automated gathering of livestock data to support the expertise and management decisions made by farmers, vets, and authorities. The high mobility of UAVs combined with a high level of autonomy, sensor-driven technologies and AI decision-making abilities can provide many advantages to farmers in exploiting instant information from every corner of a large farm. The key objectives of this research are to i) explore various drone-mounted vision-based remote sensing modalities, particularly, visual band sensing and a thermal imager, ii) develop UAV-assisted autonomous PLF technologies and ii) collect data with various parameters for the researchers to establish further advanced AI-based approaches for monitoring livestock in large farms effectively by fusing a rich set of features acquired using vision-based multi-sensor modalities. The collected data suggest that the fuse of distinctive features of livestock obtained from multiple sensor modalities can be exploited to help farmers experience better livestock management in large farms through PLF.