Estimation of Airflow in Livestock Buildings using Image Analysis

In order to evaluate a given ventilation system in a livestock building and its sensitivity to wind, presence of heat sources (e.g. livestock) etc. it is of interest to estimate flow vector fields corresponding to the airflow. By introducing particles (e.g. smoke) into the air inlets of a model of a livestock building, the airflow in a laser-illuminated plane may be visualized. Based on sequences of images recorded of this plane using a video camera, estimates of 2--D flow vectors are derived locally. The local estimates of velocity are found using a set of spatio-temporal convolution filters on the image sequence. After which the local estimates are integrated to smooth flow fields using a model that incorporates spatial smoothness. The algorithms are illustrated using a scale model of a pigs sty under isothermic conditions. It should be noted that this is an non-invasive technique (laser and camera may be placed so they do not disturb the airflow) for estimating air velocity in a 2--D plane.