Joint International Agricultural Conference 2009 location:Wageningen, Netherlands date:6-9 July 2009
Lameness is an increasing animal welfare problem. It has a negative impact on milk production, body condition and reproductive performance in dairy cows (Booth et al., 2004; Holzhauer et al., 2004). Since early identification and treatment can reduce the cost of lameness and improve animal wellbeing, more objective measures to quantify lameness need to be used (Keegan et al., 1998).
In the development of an automatic on-line tool for early lameness detection several image parameters are addressed. To avoid the need to redevelop algorithms for every image parameter used in lameness detection, we propose a model based motion scoring system.
The model of the cow is derived from an image stream. First of all the cow is extracted from the background using image processing techniques (Song et al., 2007). Now the derivation of the cow model is performed in a top-down approach: The overall position of the body of the cow in the image is detected. This is performed by fitting a rectangle onto its body. This provides us with information about the position and orientation of the body; in the next step the position and orientation of the head of the cow is detected in a similar way; and finally the positions and orientations of the individual limbs are detected.
Repeating this procedure gives us a mathematical model of the body position of the cow within each image and it provides us with an objective description of the motion of the cow. The model and motion description can be used for derivation of the image parameters needed for lameness detection.