According to MEMS Consulting, Scottish researchers have recently developed a 3D vision system to automatically detect the location of pig tails, thus alerting farmers to potential tail-biting incidents so they can prepare in advance to prevent such incidents from occurring.
According to the Scottish Agricultural College (SRUC), tail biting in growing pigs often occurs without warning, but once it breaks out, it can cause pain and even disease in the bitten pigs, resulting in economic losses for farmers, especially when the meat spoils due to infection from the tail wound. The college is currently collaborating on this research project with Innovent Technology Limited, pig supply chain partners (including Harboro, a Northeast Animal Feed company), and the Centre for Precision Innovation in Agricultural Engineering (Agri-EPI).
Furthermore, immediate tail docking at birth can effectively reduce tail biting incidents later on to some extent, but this method is not advisable and should be prohibited according to EU Council Directive 2008/120/EC. When tail biting occurs in growing pigs, there are many risk factors, and outbreaks can occur without any warning or obvious cause, making management difficult for farmers. Recent studies have shown that pig behavior changes before a destructive tail biting outbreak. Specifically, pigs lower their tails before a potential incident occurs.
To alert farmers to potential tail-biting incidents in pigs, the project team developed an automated system based on 3D cameras to detect tail positions. Each pigpen is equipped with an IFMO3D3013D camera, covering approximately one-third of the area, positioned above the feeder and pointing vertically downwards. The camera uses Time-of-Flight (ToF) technology, emitting infrared light pulses from an LED at a rate of 25 pulses per second and then recording the delay between the pulse and its arrival at each pixel. Data collected from each camera is transmitted via Ethernet cable to CompuLab's fit-PC4 fanless industrial PC, which is then connected to broadband internet for data download.
Each pigpen is also equipped with two 2D CCTV (closed-circuit television) bullet cameras from Gamut, featuring Sony effio chipsets, and mounted on the top of the pen; one monitors the entire pigpen, while the other monitors the area above the feeder where the 3D camera is located. These cameras record continuously 24 hours a day, and the video data is stored on the GeoVision computer-based CCTV system hard drive. Users can simultaneously view both 3D and 2D video images to "visually" verify the 3D data.
According to the academic paper "Automatic Early Warning of Pig Tail Biting: 3D Cameras Can Detect Tail Lowering and Prevent Tail Biting Incidents," Innovent Technology Ltd. has developed a proprietary algorithm to locate pigs and their tails. For each pig standing under the camera, the algorithm locates its tail and measures the angle of the tail relative to its body, ranging from 0 to 90 degrees. 0 degrees means the tail is drooping or tucked into the body and cannot protrude from the curve of the back or rump, while 90 degrees means the tail is upright at a 90-degree angle.
During camera setup and algorithm verification, the accuracy of detecting low-tailed pigs reached 73.9%, while its sensitivity was 88.4% and specificity was 66.8% compared to non-low-tailed pigs. This test involved eight batches of samples, each divided into 23 groups of 29 pigs, all raised under specific commercial conditions using non-tail docking. Tail biting incidents occurred in 15 groups. In such cases, immediate reinforcement of the pigpen fencing was necessary, and affected pigs were moved and treated. 3D data from the tail biting outbreak group showed that the proportion of low-tailed pig detection increased before the outbreak and decreased afterward.
“This research has met all our expectations. We can autonomously detect the posture of pig tails, and the study has proven to be an early warning system for tail biting incidents,” said Dr. Rick D'Eath, the first author from SRUC. “Our challenge now is to develop this advanced technology into a robust farm early warning system applicable to any pig farm.”
This research will be taken over by a new three-year project called TailTech, funded by £676,000 from the UK-based company Innovate. The project will reportedly collect more data from various pig farms and develop and test a prototype early warning system. Other pig supply partners will also participate.
Disclaimer: This article is a reprint. If there are any copyright issues, please contact us promptly for deletion (QQ: 2737591964). We apologize for any inconvenience.
