Computational Intelligence Applied to Monitor Bird Behaviour

Abstract

The best way to obtain relevant information about the behaviour of animals is direct observation (of individuals). However, traditional close-up observations can interfere on the behaviour, and taking biometric measurements requires the capture of individuals, which also causes stress. This paper describes an automatic motoring system for birds breeding in nest boxes. The main goal is to significantly increase the amount and quality of data acquired on bird behaviour without stressing the individuals or interfering. This system is based in an interconnected embedded sensor network, which permits sharing this valuable information with researchers all over the world through the internet. Each device of the network is a smart nest-box that allows a cross-validation of sensor information and data quality. This system has been evaluated for the specific case of a lesser kestrel breeding colony in Southern Spain. The lesser kestrel is an insectivorous migratory falcon that readily accepts nest-boxes. The system has been named HORUS and the results obtained from a year experiment demonstrate the efficiency of this approach.