Real-time bird detection based on background subtraction

Abstract

Detection of birds in air is an important problem across multiple applications including aviation safety, avian protection, and ecological science of migrant bird species. In this paper we describe a real-time detection system of birds in flight. Using a single fixed camera, our bird detection system is based on background subtraction and tracking through point correspondence. We make use of Zivkovic's background subtraction approach which includes a non-parametric model and a Gaussian mixture model that is an extension of the standard method. We append a correspondence component based on point-tracking to the background subtraction algorithm to achieve reliable bird detection. Experiments were conducted to study the detection performance using objects of different size, color and velocity. The results show efficiency and accuracy of our system in the detection of fast motion objects such as birds.