Análisis del modelo YOLO-V8 para la segmentación de salmones en jaulas de cultivo en proceso estimador de biomasa, en tiempo real

Abstract

The biomass estimation pipeline is an important part of the project to fully automate the salmon farming process, contributing to periodic surveillance to assess the fish’s health among other variables of interest. In the present work, we analyze the performance of one state-of-the-art instance segmentation model after a series of training experiments. Later, we compare its results with three other segmentation models, one of them being part of the current implementation of the biomass estimation pipeline. The results show the model’s ability to segment fish effectively, without compromising its realtime performance.