Stratoni Seahorse Colony, Greece

*Note: the installation and use of the NOUS system to monitor and observe the colony of seahorses in Stratoni is a pilot program still in testing.

In 2007 Vasilis Mentogiannis had an extraordinary encounter with seahorses in the waters of Stratoni in northern Greece. Thirteen years later, thanks to the pioneering technology that the NOUS team has developed, a world-first was accomplished – real-time video streaming of the Stratoni seahorse colony went live to the world for the very first time. Harnessing the power of the sun through an autonomous solar system, NOUS’ Stratoni installation allows us to watch the seahorses as they go about their fascinating lives inhabiting ingenious ‘seahorse hotels’. The hotels are an innovation of the Hippocampus Marine Institute and consist of protective metal frames that create safe spaces for the seahorses by keeping out predatory animals while a network of artificial ropes traverses the interior offering the favoured holding perches of the seahorses.

The NOUS Stratoni set-up has two core lines of development. First, the system establishes real-time monitoring through the information pipeline from the sensors to the cloud. Then, the platform’s customized neural network structure offers efficient and advanced automated monitoring capabilities base on training with local data and machine learning. This application of the NOUS system can be adapted to a number of species and environments. 

This constant window into the underwater world of the Stratoni seahorses provides an invaluable opportunity for scientists to observe the seahorses unimpeded behavior over great periods of time and to record significantly more unobstructed data than is obtainable through any other currently-existing method of study.

The NOUS Stratoni system set-up consists of prototype submarine housings equipped with cameras (day and night Infra-red), lights, a temperature measurement sensor, and windshield wipers for the camera lenses. Additionally, the system can be extended to include a hydrophone connected to the housing which would provide acoustic data. The whole operation is controlled by a NOUS multitasking computing unit. The network of the underwater cameras culminates in a submarine hub, which is powered by a cable from the shore or by a surface buoy (the latter with power from solar panels). The NOUS Stratoni system capacity is estimated to be up to 10 connected cameras grouped in two hubs. Additionally, the system can be equipped with a hydrological current meter for flow monitoring. Data from the NOUS Stratoni system is transferred through an internet connection, via fiber-optic and a GSM link to a server and cloud. The deployed system is also able to collect both complete weather data along with solar power data (in case of buoys).

Stratoni’s NOUS system is designed to offer constant observation capabilities for the artificial habitats created by the Hippocampus Marine Institute (HMI) for the seahorses. 

A combination of two cameras will provide close-up and wider visual information, offering the ability to compare information from the HMI artificial habitats to data collected from natural habitats such as sea grass and tube worms.

Processed data supplied by the NOUS-mounted hydrophone provides early alerts of acoustic disturbance at the sea floor caused by wave energy. Similarly, through continuous flow monitoring, the NOUS Stratoni system will automatically collect and correlate the patterns of the local flow circulation of microorganisms which contribute to the main food source of the seahorses.

The proposed NOUS Stratoni system is equipped with Artificial Intelligence capabilities that allow it to distinguish, classify, associate and perceive significant differences in measurable parameters that take place in the aquatic environment. These data can be extremely beneficial to the ongoing study of the area and are considered of scientific interest. Machine Learning and Deep Neural Networks will implement the above algorithms for intelligent information processing from images and spectrograms for sound recognition.