Élaboration d'un outil de suivi de l'état sanitaire de vaches laitières en pâture par thermographie

Abstract

Agriculture and farming are currently facing several challenges and must adapt to the current context in order to ensure their own sustainability. The switch of the customer’s behavior has an important impact on the evolution of these sectors. Indeed, the market is trending towards higher quality products that respect both animal welfare and environment. For these reasons, pastures are one of the most adapted practices but it requires optimized herd monitoring methods. The ROAD-STEP project aims to design a fully automated tool for monitoring the health and welfare of animals in the pasture. The body temperature is an established diagnostic tool in dairy farming that is suitable for that kind of monitoring. The purpose of this work is to design and evaluate an automated temperature measurement device using thermal imaging technologies which can be involved in the ROAD-STEP project. A low resolution FLIR Lepton 3.5 sensor-based device has been conceived and evaluated under controlled conditions. A test in real conditions was also carried out. These experiments show that this kind of sensor is able to detect surface temperature variations of the order of 0.5 ° C. However, further studies must be carried in order to correct the influencing factors and identify the orbital zone. The arrangement of the device on the gantry must also be reviewed to maximize the proportion of framed and sharp images. Also, a reliability problem with the Lepton 3.5 sensor has been identified. Reliability can probably be improved by optimizing the acquisition program. Otherwise, using continuously this kind of sensor over the long term might be more complicated than expected. There are other providers which sell similar sensors that may be more reliable.