Innovative Off-Body Monitoring of Dairy Calves' Vital Signs for Early Disease Detection: A Thermal Imaging Approach

Calves' early-life performance affects future productivity. Productivity is hampered when the animal gets sick. It is known that disease detection in dairy calves is challenging due to manual labor and limited sensor data. New developments allow for remote vital sign monitoring, including well-established physiological parameters (respiratory rate; RR and heart rate; HR). Potentially these vital signs support early detection without the difficulties often present with on-body techniques. We explored the potential of a thermal camera (Seek, California, USA) for RR and HR detection compared to gold standards; data obtained by Vernier and Polar belts, respectively. Six RR videos and five HR videos were analyzed from a single calf. RR was determined by tracking temperature dynamics around the nose, and HR from temperature variations from the eye's blood vessels. A custom Python script (version 3.9.0) was used to enhance video contrast, track regions of interest, and apply Fourier Transformation and Butterworth filtering to extract these vital signs. Pearson's correlation coefficients for reference and thermal video data were calculated using Python (version 3.9.0). Robust RR and HR correlations (RRR = 0.9816 and RHR= 0.941, respectively) were observed despite temperature fluctuations and calf movements. Specific postures, such as 'lying' and 'drinking', consistently exceeded one-minute durations, proving ideal for accurate measurements. Despite the thermal camera's limited sensitivity and inherent cost, our study confirms its feasibility for practical RR and HR detection.