Evaluating Deep Learning Assisted Automated Aquaculture Net Pens Inspection Using ROV

TL;DR

This paper introduces a real-time, deep learning-based system for automated inspection of aquaculture net pens using ROVs, improving accuracy and efficiency over traditional manual methods under various underwater conditions.

Contribution

The paper presents a novel deep learning approach for real-time defect detection in aquaculture nets using ROV video streams, with effective segmentation under challenging underwater conditions.

Findings

High detection accuracy in diverse underwater scenarios

Effective real-time processing on embedded platforms

Outperforms existing state-of-the-art methods

Abstract

In marine aquaculture, inspecting sea cages is an essential activity for managing both the facilities' environmental impact and the quality of the fish development process. Fish escape from fish farms into the open sea due to net damage, which can result in significant financial losses and compromise the nearby marine ecosystem. The traditional inspection system in use relies on visual inspection by expert divers or ROVs, which is not only laborious, time-consuming, and inaccurate but also largely dependent on the level of knowledge of the operator and has a poor degree of verifiability. This article presents a robotic-based automatic net defect detection system for aquaculture net pens oriented to on-ROV processing and real-time detection. The proposed system takes a video stream from an onboard camera of the ROV, employs a deep learning detector, and segments the defective part of the image from the background under different underwater conditions. The system was first tested using a set of collected images for comparison with the state-of-the-art approaches and then using the ROV inspection sequences to evaluate its effectiveness in real-world scenarios. Results show that our approach presents high levels of accuracy even for adverse scenarios and is adequate for real-time processing on embedded platforms.