Computer vision methods applied to semi-automated analysis of inter-individual distances of perching birds

Borut Kumperščak (2010) Computer vision methods applied to semi-automated analysis of inter-individual distances of perching birds. EngD thesis.

PDF Download (3068Kb)

Abstract

A system for analysis of perchnig bird flocks based on computer vision technologies and OpenCV library. Input data are high resolution photos of perching birds with uniform background. Number of birds, their sizes, and positions is computed out of these images. First, the objects are found that are reasonably assumed to be birds and segmented out of the background by means of computer vision methods. Analitic measurements are made on these objects, and results are saved to a database. Gathered data supports research in the field of modelling flocks of autonomous agents.

Actions (login required)