Detection of Chewing Strokes from Jaw Movement Signals in Dairy Cows Using a Nose-Mounted Accelerometer

Area of Science:

• Animal Science
• Agricultural Engineering
• Sensor Technology

Background:

• Chewing activity is a key indicator of dairy cow health and welfare.
• Accurate monitoring of chewing strokes is crucial for assessing rumen function and feeding behavior.
• Current methods for monitoring chewing can be labor-intensive and subjective.

Purpose of the Study:

• To evaluate a non-invasive, nose-mounted accelerometer for automated detection of chewing strokes in dairy cows.
• To compare the accuracy of different algorithm versions for chewing stroke detection.
• To assess the potential of automated chewing detection for precision livestock farming.

Main Methods:

• Utilized a nose-mounted accelerometer on 15 Holstein Friesian cows.
• Collected acceleration data and validated against manual video annotations.
• Applied a peak detection algorithm to smoothed acceleration data, analyzing raw and cleaned versions (excluding regurgitation intervals).

Main Results:

• The cleaned algorithm version demonstrated high agreement with manual annotation (ICC = 0.91).
• The cleaned version showed significantly lower error metrics (MAE: 3.67, RMSE: 4.72, MAPE: 5.64%) compared to the raw version.
• Automated detection of chewing activity using the sensor system was proven feasible and reliable.

Conclusions:

• Automated chewing stroke detection using a nose-mounted accelerometer is a reliable method.
• This technology can provide objective data for assessing rumen function, feeding behavior, and animal welfare.
• The system supports precision livestock farming by offering insights into dairy cow chewing activity.