Fish-NET: Advancing Aquaculture Management through AI-Enhanced Fish Monitoring and Tracking

This study seeks to enhance aquaculture and fishery management using artificial intelligence, focusing on Nigerian catfish farming. The methodology encompasses a sequence of steps from data collection to validation. A dataset, primarily composed of aerial imagery from catfish ponds and supplemented with additional data from the internet, formed the foundation of this research. By leveraging computer vision and deep learning techniques, the data were processed to assess the potential of the three distinct cutting-edge object detection models. Based on various evaluation metrics to gauge their effectiveness in fish detection tasks, the Faster R-CNN emerged as the optimal model, boasting a superior balance of precision and recall. This model was subsequently integrated with an object-tracking model and deployed as an application, yielding promising results in terms of fish detection and tracking. The findings in this study suggest that AI-driven tools can automate monitoring processes, significantly increasing accuracy and efficiency in resource utilization.