Development and application of an image acquisition system for characterizing sow behaviors in farrowing stalls

Abstract

Animal behavior can be an indicator of animal productivity and well-being, and thus an indicator of how animals respond to changes in their biophysical environment. This study monitored the behaviors of sows and piglets in a commercial setting utilizing an autonomous machine vision system. The objectives of this research were to: (1) implement a digital and time-of-flight depth imaging system, (2) develop a process with minimal user input to analyze the collected images, and (3) calculate the hourly and daily posture and behavior budgets of sows housed in individual farrowing stalls. Depth sensors were centered above each stall in three farrowing rooms (20 sows per room) and controlled by mini-PCs, acquiring images continuously at 0.2 FPS. Data files were transmitted via Ethernet cable to a switch, then to a 50 TB disk station for storage. Recorded image data were subsequently analyzed to quantify sow posture budgets and behaviors using a computer processing algorithm. Algorithm classifications were compared to those of trained human labelers with sow posture classified correctly >99.2% (sitting: 99.4%, standing: 99.2%, kneeling: 99.7%, lying: 99.9%). Specificity and sensitivity parameters for posture classifications were 2 >84.6%, with the exception of lower specificity for kneeling (20.5%). When lying, direction (sow lying on left or right side of body) was classified with an accuracy of 96.2%. Sows that were not lying were also labeled with a behavior, including feeding (97.0% accuracy), drinking behavior (96.8% accuracy), and other behavior (95.5% accuracy). Each non-lying behavior label had specificity >88.3% and sensitivity >77.4%. This autonomous system enables acquisition of a large amount of replicated data to evaluate the effects of changing the farrowing environment on sow behavior and potentially well-being.

Evaluation of a decision support system for precision feeding of gestating sows

“…Reducing apparent feeding motivation of gestating sows should therefore improve their welfare. In order to include components of behavior in the DSS, software analyzing video recordings have been developed and should be able to detect automatically abnormal behaviors like aggressive interactions ( Oczak et al, 2013 ; Viazzi et al, 2014 ) or sows’ posture ( Nasirahmadi et al, 2015 ; Leonard, 2019 ). Moreover, diet formulation options should be added in the model and available “in real time”.…”

Large-Scale Phenotyping of Livestock Welfare in Commercial Production Systems: A New Frontier in Animal Breeding

“…In the swine industry, farrowing is a challenging stage for both the sow (transition from gestation to farrowing and lactation), and the piglets (susceptibility to crushing, chilling, and malnutrition; Marchant et al, 2001;Johnson and Marchant-Forde, 2009). Aiming to identify solutions to these issues, Leonard et al (2019) monitored behavioral activities of sows and piglets in a commercial setting utilizing an autonomous machine vision system. A digital and ToF depth imaging system was implemented and a process with minimal user input to analyze the collected images was developed to calculate the hourly and daily posture and behavior activities of sows housed in individual farrowing crates.…”

A Systematic Review on Validated Precision Livestock Farming Technologies for Pig Production and Its Potential to Assess Animal Welfare

“…Drinking patterns can be monitored using video analysis for evaluating visits to the drinker and contact time (94,95), and for distinguishing drinking from drinker-playing behavior (96). Cameras for the identification of behavior of sows were used for identifying feeding and drinking behavior in the farrowing crate (37,73,74), as well as in group-housed sows (97).…”

“…Internal validation of the vision-based technologies in many cases reported very promising results with accuracy above 95% [e.g., (30,33,58,73,74,93,95,97,107,(145)(146)(147)]. However, none of the outstanding performance results for vision-based monitoring have been confirmed by external validation.…”

“…However, none of the outstanding performance results for vision-based monitoring have been confirmed by external validation. Image-analysis has been used for assessing sows' postures, such as standing, lying, and sitting (73), evaluating lying patterns in group-housed pigs responding to thermal conditions of the pen (76, 80), detecting animals' location (32,84), distinguishing drinking and drinkerplaying behavior (96), identifying feeding behavior (37,73,74,97), recognizing aggression events (100-103) and tail biting (50), estimating body weight (36,118,119), and detecting African swine fever (31).…”