Internet of things-based smart fish farming: Application of smart sensors and computer vision to provide real-time monitoring and diagnosis in aquaculture

Abstract

The frequent occurrence of disease outbreaks in fish farming presents a significant challenge, leading to substantial economic losses and threatening food security, thus hindering the progress toward sustainable development goals (SDGs). In aquaculture, disease prevention relies on early detection of changes in water quality, abnormal fish behavior, and physical deformities, tasks typically handled by skilled fisheries experts, who are in short supply in Nigeria. Traditional manual disease detection methods are often costly and unreliable. This study proposes a computer vision-based solution utilizing Faster Region-based Convolutional Neural Network (FasterR-CNN) with Detectron2 for improved disease detection in fish farming. A dataset of 500 images was collected, pre-processed, and divided into training (70%), validation (15%), and testing (15%) sets. Three Faster R-CNN models (X101, R100, and R50) were trained and evaluated, with the X101 model achieving the highest accuracy of 98%. The results underscore the potential of deep learning techniques for accurate and efficient disease detection, offering a scalable solution to enhance fish health management. This approach provides a reliable and cost-effective alternative to traditional methods, contributing to the sustainability and growth of the aquaculture industry while addressing the need for timely interventions in fish disease control.