Automated method for measuring body size parameters of live pigs based on non-rigid registration of point clouds

With the increasing demand of human and the development of automation technology, there is an urgent need for automated livestock body measurement methods. An object of this paper is to investigate an automated method for measuring body size parameters of live pigs based on non-rigid registration of point clouds. We choose Xtion pro camera as point cloud acquisition equipment to obtain RGB-D data from the upper left and upper right. Pig body point clouds are extracted from the acquired data after preprocessing. We perform coarse registration to the measured point clouds and the template point cloud so that the two point clouds are aligned. Moreover, non-rigid registration methods are used in order to make the shape of the measured data fit with the template data. Finally, the different body size parameters are estimated according to the landmarks that marked on the template point cloud in advance. The results have shown that this method is reliable for measuring pig body size parameters compared to manually measured values, especially in measuring the hip height, shoulder width, and hip width, in which the average relative errors are 2.92%, 5.78%, and 4.83%.