Improving Farm Efficiency and Animal Health, Welfare and Productivity
The current lack of accurate predictors for breeding sow lameness, body condition, and parity potential does not allow farms to reach their maximum revenues as animals either die prematurely or are culled from the herd. Most data on sows is collected by farmhands and although there are apps that enable the recording of data collected, the device itself is often not automated. The use of calipers, a tool widely used to measure sow condition and adjust feed rations, is a manual and labor-intensive operation.
Although a lot of data is collected regarding sow health and performance, the general consensus is that the data is used as a historic record but is not used to predict outcomes.
The prediction of outcomes is the essence of Motion Grazer and the use of its tools will drive improved productivity while reducing costs.
Artificial intelligence (AI) has enabled a new era of computer vision where complex sensing tasks no longer require hand-crafted models. Instead, convolutional neural networks trained on large datasets can achieve far more accurate and robust solutions than hand-crafted approaches. This has led to a revolution in performance in areas such as object recognition, person tracking, and 3D shape estimation. Outcomes of using this technology have enabled specific tracking of human joint location as well as evaluation of gait.
Motion Grazer AI (Motion Grazer) is taking these same tools to the farm where it will initially analyze pig shape and gait in conjunction with tracking joint location. Such analyses will be used to predict lameness and body condition of female breeding pigs (sows) to provide quantitative rather than subjective data to identify which sows to be maintained for breeding and those to be culled.
Sow Shape & Condition Determined From Depth Analysis
Sow Gait From Tracked 3D Joint Locations
Extensive Data Collection and Machine Learning
Sow Health and Productivity Prediction
Above: Stationary sows shown are for illustrative purposes. Animals are in motion when scanned.
- Low cost unit
- Capture changes in body composition to make feed adjustments
- Save labor time and person power
- Enhance sow health and welfare
- Provides quantitative tool to refine and improve economic decisions
- Improve efficiency and increase productivity
- Evaluates gait and animal condition (snout to tail)
- Data collected without distraction or prejudice