Physically Separated but Socially Connected: Physiological Stress Evaluation in Pigs using Multimodal Sensor Data
In our current systems of housing livestock animals, pigs experience stress through multifactorial situations. In order to objectively assess the stress the animals experience, digital livestock farming tools and platforms comes quite handy. The aim of this study was to investigate the usage of thermographic imaging platforms and heart rate sensor wearables in the assessment of the pig welfare while housing as part of a 4 month feeding experiment study and during transportation to the slaughterhouse. Heart rate measurement sensor belts were attached to five pigs, and the same pigs were simultaneously filmed with a thermographic camera while housed in pairs as well as housed separated, both before and after feeding, resulting in four treatment groups, paired/before feeding, isolated/before feeding, isolated/after feeding and paired/after feeding. We also employed the heart rate monitors on the pigs during transportation experiments to collect vital data as part of real-time continuous monitoring of welfare indicators.
During transport, the heart rate fluctuated significantly different events, indicating that the pigs experienced stress. A difference in skin temperature between pigs as measured by thermal camera data clearly indicated the varying stress levels to the external perturbations. Measuring heart rate in pigs can give valuable insight for enhancing the pig welfare, however, the method to measure the heart rate of pigs as a wearable device while housed socially is still a hurdle that needs to be overcome. The use of a thermographic camera to look at body surface temperature showed promising results, however, hurdles associated to big data analysis and extraction of thermal scores at specific body regions of the pig need to be overcome before the actual use in production systems. Rectal temperature and salivary cortisol measurements were obtained from the pigs during all the treatments to biochemically and physiologically correlate the measurements to the multimodal sensor data. In addition to the heart rate sensor, we also acquired 3-axis accelerometer data to explore the gait and resting rhythms of the pigs before and after feeding at various intervals.
The large volume of multi-modal data obtained as part of this longitudinal study was analyzed using Artificial Intelligence enabled machine learning approaches in creating critical insights for development of decision support and recommendation systems for the animal caretaker and the pig industry sector personnel.
Correlation between Rectal Temperature and the skin temperature thermal scores as measured by the thermal camera from the pigs
Wearable Biometric Sensor Reveal the Secrets of Pig Behaviour by measuring 3 vital data simultaneously (Heart Rate, Respiration Rate and Acceleration Info)