Vertical back movement of cows during locomotion: detecting lameness with a simple image processing technique

Area of Science:

• Animal Science
• Computer Vision
• Agricultural Engineering

Background:

• Lameness is a significant welfare and economic issue in dairy farming.
• Accurate and early detection of lameness is crucial for timely intervention.
• Existing detection methods can be labor-intensive or require specialized equipment.

Purpose of the Study:

• To develop and validate a simple, automatic image processing technique for lameness detection in dairy cows under farm conditions.
• To assess the diagnostic accuracy and correlation of the proposed system with visual lameness scoring.

Main Methods:

• Video capture of 75 dairy cows on a designated walking path.
• Image analysis of back arch posture to derive an automatic lameness score (ALDS).
• Calibration with 12 cows and validation with 63 cows against a reference lameness score (RLS).

Main Results:

• The ALDS achieved 100% sensitivity and specificity for binary lameness detection (lame/non-lame).
• Perfect agreement (ρc = 1) and strong correlation (r = 1) were observed for binary scores.
• Strong agreement (ρc = 0.885) and high correlation (r = 0.840) were found for ordinal lameness severity scores (LS1-LS5).

Conclusions:

• The proposed automatic lameness detection system demonstrates high diagnostic accuracy and reliability.
• This image processing technique shows potential as a non-invasive, farm-friendly tool for dairy cow lameness monitoring.
• The method could offer a competitive alternative to existing vision-based lameness detection systems.