Detection, identification and posture recognition of cattle with satellites, aerial photography and UAVs using deep learning techniques

To obtain specific information about cattle in extensive production systems, the usual labor intensive work done by farmer to find and visit herds large pastures can be replaced using UAVs. UAVs are capable of assessing traits cows, like distinguishing individuals postures. Although these detection cattle, do not represent resilience efficiency directly, may contain associated resilience. We performed a feasibility study remotely sensed imagery (using datasets from satellites, manned aircrafts, UAVs), deep learning techniques detect, count, identify characterize posture individual cows grassland systems. With techniques, we focused on : (1) automatic locations animal counting; (2) cow postures standing, grazing or lying; (3) identification. Data were collected during three field trials Netherlands Poland. Artificial Intelligence was used classify objects (cattle) drone imagery. Classification accuracies >95% obtained for detecting cows. Accuracies ~91% identifying ~88% These results make camera-mounted drones promising new technology monitoring beef