Struck based infrared flying bird tracking experiments

Bird strikes present a huge risk for air vehicles, especially since traditional airport bird surveillance is mainly dependent on inefficient human observation. Recently, multiple groups have proposed bird monitoring using computer vision to detect birds, determine bird flying trajectories, and predict aircraft takeoff delays. However, the characteristics of bird flight using imagery which should be handled by visual tracking algorithm are not well known. Therefore, we perform bird flight tracking experiments using a robust approach, Struck, to obtain useful clues and recommend feature analysis. We also develop a Struck-scale method that uses the power of Struck while accounting for different image sizes through multiple scale sampling adaption. Using four types of features and four types of kernel functions, a performance comparison is demonstrated on a real infrared bird flight dataset. Through experiments, we conclude six characteristics of the bird tracking. The general analysis can be used in the bird tracking methods for airport safety, remote bird monitoring, and urban bird population studies.