AlyVesi -project using 5G and machine learning in real time

News 2019-08-30 at 13:28
Photo Otto Uotila.

Video from a drone, computer vision and superfast 5G connections were harnessed for monitoring the Baltic Sea in a trial conducted in Kirkkonummi. In the trial Nokia, Nordkapp, the AlyVesi-project from Finnish Environment Institute SYKE, Telia and Vaisala developed new tools for SYKE researchers for monitoring the environment.

The AlyVesi -project is interested to develop use of 5G for SYKE for also situational flood awareness, monitoring in hydrology, flood risk management and preparedness, environmental monitoring and biomass assessments in water bodies. In the AlyVesi –project, which is part of the Government Blue bioeconomy program, machine learning is widely used in water sector. Fast data transfer rate, combined with a beyond visual line of sight drone operations and machine learning provides efficiency for real time operations.

"Until today we have mapped large areas of invasive species in Finland and Estonia. As more often multispectral images are used higher data transfer rates are more than welcome. Today we can insure data quality in real time and get better quality with minimum effort, no need for revisits" says AlyVesi project manager Jari Silander from SYKE's freshwater center.

In the 5G trial, the blue-green algae situation was monitored with a drone and computer vision. The camera and sensor -equipped drone was flown over the Baltic Sea, and the high-resolution video was transmitted over 5G for real-time analysis. A license plate number is easily recognizable from a drone form 700 meters away from the target. This is especially helpful when mapping vast areas of interesting species from drones.

“Blue-green algae monitoring is based on multiple sources of information, including satellite imagery and automated chlorophyll measurements from ferries sailing the Baltic Sea. This data is combined with local visual observations made at the shoreline. In the trial in Kirkkonummi, the drones operated over a wide area outside the line of sight, and the information was transferred in real time to computer vision. Under good conditions, computer vision detected blue-green algae with over 90 percent accuracy,” says Jari Silander of SYKE.

More information

Batteries weight over half of the weight of the drone. © Jari Silander.
In the city of Kirkkonummi a drone fleet was used to map interesting targets. © Jari Silander.




Target group: