부산대학교 도서관

Since migrating birds can spread the Avian Influenza (AI), AI expansion can be reduced or blocked by establishing a comprehensive surveillance system on the bird migration pattern by observing and tracking their migration paths and habitats. It is necessary to develop technological approach to observe and track the migrating birds. We propose a method for detecting migratory birds toward observation by using Haarlike features and a feature map. The process of using Haar-like feature can classify groups between adjacent dark and bright from a river scene with floating migratory birds. This paper focuses on success rate of detection of migratory bird or group of migratory birds by combining optimization algorithm with a geometrical feature map. We also present the experimental results on bird detection, by comparing previously achieved results.