高速ビデオ画像を用いた遠距離鳥類識別アルゴリズム

抄録

Monitoring avian activities is an essential procedure for bird behavior research and pest damage management in agriculture. Existing bird detection and identification methods work well at near distances but they are rather ineffective at long distances. We proposed a novel avian identification algorithm based on high-frame-rate (HFR) video that is able to detect flying birds at a distance above 50 meters. In order to identify birds that fly freely over a long distance and in a wide range, we use the wing shape, flapping, and movement of the center of gravity of birds.