Sow and Piglet Behavior Characterization Using Visual Observation, Sensor Detection, and Video Recording

Authors: KIM, JUN HO - Purdue University; NI, JI-QIN - Purdue University; OGUNDARE, WONDERS - Purdue University; SCHINCKEL, ALLAN - Purdue University; MINOR, RADIAH - North Carolina Agricultural And Technical State University; Johnson, Jay; CASEY, THERESA - Purdue University

Submitted to: American Journal of Applied Sciences
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 3/7/2025
Publication Date: 3/11/2025
Citation: Kim, J., Ni, J., Ogundare, W., Schinckel, A.P., Minor, R., Johnson, J.S., Casey, T.M. 2025. Sow and Piglet Behavior Characterization Using Visual Observation, Sensor Detection, and Video Recording. Biosystems Engineering. 15(6), 3018. https://doi.org/10.3390/app15063018.
DOI: https://doi.org/10.3390/app15063018

Interpretive Summary: Non-invasive and automatic detection of pig behavior without the use of cameras may be a valuable tool to assess pig welfare in commercial production systems. As such, the study objective was to compare the accuracy of direct observation, non-video sensor detection, and video recording to evaluate pig behavior as an indicator of welfare status under heat stress conditions. In the study, half of the sows were subjected to heat stress and the other half were kept cool. It was determined that heat stressed sows under heat stress spent more time drinking, sitting, and lying down but less time eating, standing, nursing, and sleeping, while their piglets slept more but nursed less. Behaviors detected by the non-video sensors generally agreed visually observed behaviors, particularly for active movements like eating and standing, but were less accurate for passive behaviors such as lying down and sleeping. Video recordings confirmed the sensors' ability to capture subtle movements, providing detailed insights into pig behavior. The combination of visual observation and non-video sensor technology offers promise for better understanding and improving pig welfare under commercial production conditions.

Technical Abstract: Animal behavior is an indicator of stress, disease, and overall well-being for sows and piglets. This study evaluated three methods for characterizing swine and piglet behaviors: visual observation, sensor detection, and video monitoring. An experiment was conducted in 18 swine crates in a research building. The sows were randomly assigned to either a natural heat stress (n=9) treatment or cooling (n=9) treatment using water cooling pads. Video monitoring was performed using Zmodo wireless cameras for continuous behavior recording. Five sow behaviors and three piglet behaviors were recorded through direct visual observation every 5 minutes over a continuous 48-hour period. The sensors detection was performed continuously using one passive infrared detector in each crate. Visual observation revealed distinct pig circadian rhythms and different patterns between the treatments. The sows had peak times in eating, standing, and drinking approximately from 05:00 to 12:00 and 16:00 from to 22:00. Sows under heat-stress spent 27.4%, 36.1%, and 49.3% more time drinking, sitting, and laying, respectively, but less time eating, standing, nursing, and sleeping. Piglets under heat-stress spent 9.0% more time sleeping but 17.3% less time nursing. The sensor outputs and pig moving behaviors (i.e., sow eating + standing + drinking + sitting + piglet walking) showed good correlation coefficients (' = 0.81 for heat-stress and ' = 0.74 for cooling). In contrast, the sensor outputs and sow on-body behaviors (i.e., sow laying + nursing + sleeping), which involved minimal movements, had clear negative correlations (' = -0.77 for heat-stress and ' = -0.56 for cooling). Video recording analysis validated the response and sensitivity of the sensor, which could quickly capture changes in pig behaviors, including subtle movements such as lifting ear, moving tail, shaking head, and sniffing, and provide behavioral information about the nuanced pig movements.