Multi-species Seagrass Detection and Classification from Underwater Images

TL;DR

This paper presents a deep learning-based system for detecting and classifying multiple seagrass species in underwater images, achieving high accuracy and reducing manual labeling efforts.

Contribution

It introduces a novel multi-species seagrass detector and classifier using deep convolutional neural networks, along with a semi-automatic labeling method and publicly available dataset.

Findings

Achieved 92.4% overall accuracy in seagrass classification

Developed a semi-automatic labeling approach to reduce manual effort

Released dataset, code, and models for reproducibility

Abstract

Underwater surveys conducted using divers or robots equipped with customized camera payloads can generate a large number of images. Manual review of these images to extract ecological data is prohibitive in terms of time and cost, thus providing strong incentive to automate this process using machine learning solutions. In this paper, we introduce a multi-species detector and classifier for seagrasses based on a deep convolutional neural network (achieved an overall accuracy of 92.4%). We also introduce a simple method to semi-automatically label image patches and therefore minimize manual labelling requirement. We describe and release publicly the dataset collected in this study as well as the code and pre-trained models to replicate our experiments at: https://github.com/csiro-robotics/deepseagrass