Dynamic small area visual target tracking discrimination algorithm and animal welfare evaluation applications

Abstract

Automatic monitoring and evaluation of cattle welfare status in smart pastures requires tracking and identification of the target cattle's ear area and morphology based on video images. Most of the traditional methods use contact detection, which is somewhat invasive and easy to cause cattle stress reaction. In this paper, we design a dynamic small area tracking discrimination algorithm based on the marker less posture estimation method, which includes the tracking matching model of key points, the relative fluctuation model of cattle ear and the fluctuation behavior evaluation model in turn, and finally realize the automatic recognition of motion cattle ear area and morphological features. Through simulation experiments, the effectiveness and feasibility of the method are demonstrated.