Detection, Localization and Classification of Fish and Fish Species in Poor Conditions using Convolutional Neural Networks

Jesper Haahr Christensen; Lars Valdemar Mogensen; Roberto Galeazzi; Jens Christian Andersen

doi:10.1109/AUV.2018.8729798

Detection, Localization and Classification of Fish and Fish Species in Poor Conditions using Convolutional Neural Networks

Jesper Haahr Christensen, Lars Valdemar Mogensen, Roberto Galeazzi, Jens Christian Andersen

Atlas Maridan ApS

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

Abstract

In this work the initial steps towards a system capable of parametrising fish schools in underwater images are presented. For this purpose a deep convolutional neural network called Optical Fish Detection Network (OFDNet) is introduced. This is based on state-of-the-art deep learning object detection architectures and carries out the task of fish detection, localization and species classification using visual data obtained by underwater cameras. This work is focused towards applications in the poorly conditioned North and Baltic Sea and is initially developed for the purpose of recognizing herring and mackerel. Based on experiments on a dataset obtained at sea, OFDNet is shown to successfully detect 66.7% of the fish included and furthermore classify 89.7% of these correctly.

Original language	English
Title of host publication	Proceedings of 2018 IEEE/OES Autonomous Underwater Vehicle Workshop (AUV)
Number of pages	6
Publisher	IEEE
Publication date	2019
Pages	1-6
ISBN (Electronic)	978-1-7281-0253-5
DOIs	https://doi.org/10.1109/AUV.2018.8729798
Publication status	Published - 2019
Event	2018 IEEE OES Autonomous Underwater Vehicle Symposium - Rectory Building, University of Porto, Porto, Portugal Duration: 6 Nov 2019 → 9 Nov 2019 Conference number: 13 https://auv2018.lsts.pt/

Workshop

Workshop	2018 IEEE OES Autonomous Underwater Vehicle Symposium
Number	13
Location	Rectory Building, University of Porto
Country/Territory	Portugal
City	Porto
Period	06/11/2019 → 09/11/2019
Internet address	https://auv2018.lsts.pt/

Keywords

Artificial intelligence
Deep learning
Convolutional neural networks
Object detection
Fish detection

Access to Document

10.1109/AUV.2018.8729798

OpenUrl availability

Full text

Cite this

@inproceedings{85b00f8849ba47628839b23dbf14cca1,

title = "Detection, Localization and Classification of Fish and Fish Species in Poor Conditions using Convolutional Neural Networks",

abstract = "In this work the initial steps towards a system capable of parametrising fish schools in underwater images are presented. For this purpose a deep convolutional neural network called Optical Fish Detection Network (OFDNet) is introduced. This is based on state-of-the-art deep learning object detection architectures and carries out the task of fish detection, localization and species classification using visual data obtained by underwater cameras. This work is focused towards applications in the poorly conditioned North and Baltic Sea and is initially developed for the purpose of recognizing herring and mackerel. Based on experiments on a dataset obtained at sea, OFDNet is shown to successfully detect 66.7% of the fish included and furthermore classify 89.7% of these correctly.",

keywords = "Artificial intelligence, Deep learning, Convolutional neural networks, Object detection, Fish detection",

author = "Christensen, {Jesper Haahr} and Mogensen, {Lars Valdemar} and Roberto Galeazzi and Andersen, {Jens Christian}",

year = "2019",

doi = "10.1109/AUV.2018.8729798",

language = "English",

pages = "1--6",

booktitle = "Proceedings of 2018 IEEE/OES Autonomous Underwater Vehicle Workshop (AUV)",

publisher = "IEEE",

address = "United States",

note = "2018 IEEE OES Autonomous Underwater Vehicle Symposium, AUV 2018 ; Conference date: 06-11-2019 Through 09-11-2019",

url = "https://auv2018.lsts.pt/",

}

Christensen, JH, Mogensen, LV, Galeazzi, R & Andersen, JC 2019, Detection, Localization and Classification of Fish and Fish Species in Poor Conditions using Convolutional Neural Networks. in Proceedings of 2018 IEEE/OES Autonomous Underwater Vehicle Workshop (AUV). IEEE, pp. 1-6, 2018 IEEE OES Autonomous Underwater Vehicle Symposium, Porto, Portugal, 06/11/2019. https://doi.org/10.1109/AUV.2018.8729798

Detection, Localization and Classification of Fish and Fish Species in Poor Conditions using Convolutional Neural Networks. / Christensen, Jesper Haahr; Mogensen, Lars Valdemar; Galeazzi, Roberto et al.
Proceedings of 2018 IEEE/OES Autonomous Underwater Vehicle Workshop (AUV). IEEE, 2019. p. 1-6.

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

TY - GEN

T1 - Detection, Localization and Classification of Fish and Fish Species in Poor Conditions using Convolutional Neural Networks

AU - Christensen, Jesper Haahr

AU - Mogensen, Lars Valdemar

AU - Galeazzi, Roberto

AU - Andersen, Jens Christian

N1 - Conference code: 13

PY - 2019

Y1 - 2019

N2 - In this work the initial steps towards a system capable of parametrising fish schools in underwater images are presented. For this purpose a deep convolutional neural network called Optical Fish Detection Network (OFDNet) is introduced. This is based on state-of-the-art deep learning object detection architectures and carries out the task of fish detection, localization and species classification using visual data obtained by underwater cameras. This work is focused towards applications in the poorly conditioned North and Baltic Sea and is initially developed for the purpose of recognizing herring and mackerel. Based on experiments on a dataset obtained at sea, OFDNet is shown to successfully detect 66.7% of the fish included and furthermore classify 89.7% of these correctly.

AB - In this work the initial steps towards a system capable of parametrising fish schools in underwater images are presented. For this purpose a deep convolutional neural network called Optical Fish Detection Network (OFDNet) is introduced. This is based on state-of-the-art deep learning object detection architectures and carries out the task of fish detection, localization and species classification using visual data obtained by underwater cameras. This work is focused towards applications in the poorly conditioned North and Baltic Sea and is initially developed for the purpose of recognizing herring and mackerel. Based on experiments on a dataset obtained at sea, OFDNet is shown to successfully detect 66.7% of the fish included and furthermore classify 89.7% of these correctly.

KW - Artificial intelligence

KW - Deep learning

KW - Convolutional neural networks

KW - Object detection

KW - Fish detection

U2 - 10.1109/AUV.2018.8729798

DO - 10.1109/AUV.2018.8729798

M3 - Article in proceedings

SP - 1

EP - 6

BT - Proceedings of 2018 IEEE/OES Autonomous Underwater Vehicle Workshop (AUV)

PB - IEEE

T2 - 2018 IEEE OES Autonomous Underwater Vehicle Symposium

Y2 - 6 November 2019 through 9 November 2019

ER -

Detection, Localization and Classification of Fish and Fish Species in Poor Conditions using Convolutional Neural Networks

Abstract

Workshop

Keywords

Access to Document

OpenUrl availability

Fingerprint

Cite this