The Development of WahooBay's AI Fish Classification Model

Spring 39

Overview

Identifying and monitoring fish species in underwater environments poses several challenges due to varying lighting conditions, water clarity, and complex backgrounds. Current solutions are predominantly designed for above land applications like pedestrian and vehicle detection, and do not address these challenges. The Wahoo Bay AI Fish Classification Model aims to address these limitations by developing a robust, lightweight machine learning model with the ability to detect and track fish species in real-time. Using a convolutional neural network (CNN) trained for underwater classification using our own annotated dataset from marine life found at Wahoo Bay. This approach will provide an accessible underwater experience to enhance our understanding of underwater biodiversity and raise awareness for ocean and reef conservation. This system is an accessible and engaging way to explore marine life for researchers, educators, park visitors, and marine enthusiasts.

 

Community Benefit

The project will enhance marine conservation efforts by providing researchers with a tool for identifying fish species in underwater environments. By overcoming challenges like water clarity, this technology will improve our understanding of underwater biodiversity and support ocean conservation efforts, fostering public awareness and protection of marine life.

 

Team Members

 

Sponsored By

Wahoo Bay, Center for Connected Autonomy and AI, I-SENSE