Automatic spawning detection in oysters: a fault detection approach

Hafiz Ahmed (1) , Rosane Ushirobira (1) , Denis Efimov (1) , Damien Tran (2) , Mohamedou Sow (2) , Jean-Charles Massabuau (2)

1 NON-A - Non-Asymptotic estimation for online systems
2 EPOC - Environnements et Paléoenvironnements OCéaniques

Using measurements of valve activity in a population of bivalves under natural environmental condition (16 oysters in the Bay of Arcachon, France), an algorithm for the automatic detection of spawning period of oysters is proposed. The algorithm is based on the fault detection approach and it works through the estimation of velocity of valves movement activity, which can be obtained by calculating the time derivative of the valves distance. A summarized description on the method used for the derivative estimation is provided, followed by the associated signal processing and decision making algorithm to determine spawning from the velocity signal. A protection from false spawning detection is also considered by analyzing the synchronicity in spawning. Through this study, it is shown that spawning in a population of oysters living in their natural habitat (i.e. in the sea) can be automatically detected without any human expertise, like visual analysis.

Mots clés: signal processing; spawning detection; oysters; IFault detection

Domaines: Automatique / Robotique

Hafiz Ahmed, Rosane Ushirobira, Denis Efimov, Damien Tran, Mohamedou Sow, et al.. Automatic spawning detection in oysters: a fault detection approach. Proc. ECC'15, Jul 2015, Linz, Austria. ⟨hal-01140324⟩