Torfinn Solvang and Andreas Hagemann
Machine vision represents an accurate and easily verifiable method for observing live organisms and this technology is constantly evolving in terms of accessibility and cost. Motivated by the complexity of observing small-sized aquatic organisms in experimental systems, and the difficulties related to real-time observation, sampling and counting without interfering with the organisms, we here present a new method for observing behaviour and dispersion of non-sessile zooplankton organisms using a custom-made tank with an associated machine vision system. The system was used in an experiment where the aim was to assess the effect of sound and ultrasound on the phototactic behaviour of copepodite stages of the salmon louse (Lepeophtheirus salmonis). The experimental set-up is described, including a triangular test tank designed to create a sound pressure gradient, a mechanized camera movement system and a vision system with dedicated image processing software.