The presence of behavioural rhythms has been studied in organisms of coastal areas in relation to circadian and tidal cycles, but their presence in benthic fauna inhabiting dark deep-sea regions remains largely unknown. Cabled video-observatories allow the study of these activity rhythms via the acquisition of pictures or footages over extended periods of time. In this work, we present the preliminary steps in the automation of biological data extraction for the determination of deep-sea fauna activity rhythms with TEMPO-mini (NEPTUNE; Canada). Automated analyses of tube worm behaviour were carried out with the Hough transform algorithm. Some different testing parameters were applied to the same image with siboglinid tubes. Tube openings identifications showed to be difficult, since circle placing was in some cases attributed to animals. The future step of automation will be to run the Hough transform algorithm within sub Region of Interests were tube identifications is the most efficient as we identified with this preliminary screening. Then, we will focus on each singe individual producing time series in terms of opening identification (as marker of moments of animals’ retractions) per unit of time (e.g. 10 min).