Early April, we took our friend "Wally-B" to the Sydney Fish market to train the AI model thanks to a great team from the NSW dpi. Wally-B (picture below) is a little structure equipped with a monitor and a camera linked to an AI made by Sam Aubin from the IN2ITION project. This AI's goal is to identify fish species and estimate length using a stereo camera. This would help gather data passively at boat ramps on the fish caught by recreational fishers. The aim of this project is research and will not interfere with fishers practices.
To train the model, we need to show different fish species to the AI at different angles, lighting conditions and distances to the camera. Once, we collect enough data we will put a smaller version of Wally-B into deployment at a fish cleaning table.
Kommentarer