Phytoplankton Detection using Faster R-CNN for Greenwater Quality Monitoring

Phytoplankton are microscopic marine algae that are crucial to freshwater ecosystems. These single-celled organisms are the base of aquatic food webs. They are too small to be detected in detection systems developed for greenwater quality monitoring. Monitoring the phytoplankton level in a green water pond provides information on pond inhabitants by observing their growth and well-being. We created a system for Phytoplankton detection using the object detection algorithm Faster R-CNN. A phytoplankton dataset was gathered to train the model through Python and to be run on a Raspberry Pi 4B. A 5-inch LCD was used to display the program's graphic user interface (GUI), showing the output of the captured image. As a result, a total accuracy of 90% was achieved on ten trials on a Raspberry Pi camera’s captured image, where the result of 9 trials showed little difference from actual phytoplankton counts.