Developing a quantitative method for objectively evaluating the behavior of laying hens in furnished cages
J. Ceunen1, E. Struelens2, A. Janssen3, F. Tuyttens2, K. de Baere3, J. Zoons3, B. Sonck2, E. Vranken1 and D. Berckmans1
1Laboratory for Agricultural Research, Catholic University of Leuven, Leuven, Belgium
2Agricultural Research Centre, Merelbeke, Belgium
3Provincial Agriculture Service, Antwerp, Belgium
Next to production, physiology and health, behavior is an important characteristic to consider with respect to animal welfare when evaluating novel housing systems. Behavioral characteristics are usually measured by visual observation, which is time consuming and prone to human error. Automated surveillance, by means of cameras and image-processing techniques, has the ability to generate data providing an objective measure of behavior, without disturbing the animals.
The general purpose of this study is to develop an image-processing technique to quantify the behavior of laying hens. At present, the imaging system under development is able to identify three different postures (standing, laying and pecking), by analyzing contour plots of hens and then classifying these plots.
In a series of experiments, 642 images of a single hen in a furnished cage were analysed. A model-based segmentation technique (GVF Snake's) was used to calculate the contour of the laying hen automatically. A Fourier descriptor was used to represent the contour, which was visualized by Multi Dimensional Scaling techniques in a 2-D space. The postures of the laying hen (standing, laying and pecking) formed three clearly separate clusters, with an appropriate Kruskal stress of less than 0.1 in all cases. Using these techniques, it was possible to successfully cluster the 642 images in the database according to the hen's posture.
Paper presented at Measuring Behavior 2002 , 4th International Conference on Methods and Techniques in Behavioral Research, 27-30 August 2002, Amsterdam, The Netherlands