Environmental monitoring often needs to be done in challenging environments. The FishID platform can automate analysis of imagery with poor water clarity and where other forms of monitoring are difficult. These are some of the projects the FishID team are working on.

Acoustic monitoring

Underwater acoustic cameras that use sound to detect objects have become a powerful tool for monitoring at night and in turbid waters. However, acoustic cameras, like other cameras, produce large volumes of imagery that need to be analysed.

Here we show that FishID software reliably detects and counts fish in imagery from acoustic cameras – especially using their acoustic shadows. This can substantially reduce the time, effort, and cost of monitoring with acoustic cameras.

Learn more about automating fish detection using shadows in acoustic camera images.

Turbidity removal

Turbid environments are common in coastal habitats and the wide range of environmental factors that can impact turbidity means that it is often difficult to choose and plan the best time to sample with underwater cameras.

In collaboration with Derya Akkaynak, from Haifa University, we use colour-correction algorithms to ‘clear’ the water, and are testing how much this improves the accuracy of automatic predictions in turbid environments.

Check back here for updates on this project soon.

Infrared

Infrared cameras allow the recording of animal abundances and behaviour at night, without animals having any idea that they are being observed. The FishID team has begun using infrared for sampling at night-time in clear waters, to detect species of interest that can be missed during daytime sampling.

We are working towards automating data collected on infrared cameras to further improve the capabilities of environmental monitoring.

Check back here for updates on this project soon.