E-Thesis 197 views
Using machine vision and behavioural indicators to monitor cleaner fish welfare in the Atlantic salmon (Salmo salar) aquaculture industry / ISLA MONAGHAN
Swansea University Author: ISLA MONAGHAN
Abstract
The aim of this thesis was to explore how machine vision and behaviour indicators could be used to monitor welfare of Ballan wrasse (Labrus bergyltus) and Lumpfish (Cyclopterus lumpus), which are used as cleaner fish in the Atlantic salmon industry. This thesis is divided into 4 chapters. Chapter 1...
| Published: |
Swansea University, Wales, UK
2024
|
|---|---|
| Institution: | Swansea University |
| Degree level: | Master of Research |
| Degree name: | MSc by Research |
| Supervisor: | De Leaniz, C. G., and De Olmo, S. C. |
| URI: | https://cronfa.swan.ac.uk/Record/cronfa66199 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| first_indexed |
2024-04-25T14:25:15Z |
|---|---|
| last_indexed |
2024-04-25T14:25:15Z |
| id |
cronfa66199 |
| recordtype |
RisThesis |
| fullrecord |
<?xml version="1.0" encoding="utf-8"?><rfc1807 xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:xsd="http://www.w3.org/2001/XMLSchema"><bib-version>v2</bib-version><id>66199</id><entry>2024-04-25</entry><title>Using machine vision and behavioural indicators to monitor cleaner fish welfare in the Atlantic salmon (Salmo salar) aquaculture industry</title><swanseaauthors><author><sid>fd52564ded9420c92e1edcb4d81ff39d</sid><firstname>ISLA</firstname><surname>MONAGHAN</surname><name>ISLA MONAGHAN</name><active>true</active><ethesisStudent>false</ethesisStudent></author></swanseaauthors><date>2024-04-25</date><abstract>The aim of this thesis was to explore how machine vision and behaviour indicators could be used to monitor welfare of Ballan wrasse (Labrus bergyltus) and Lumpfish (Cyclopterus lumpus), which are used as cleaner fish in the Atlantic salmon industry. This thesis is divided into 4 chapters. Chapter 1 is a systematic literature review on the use of artificial intelligence (AI) in the aquaculture industry. The review looks into examples of different uses for AI in the industry, including intelligent feeding, monitoring of water quality, behaviour and disease, and assessment of body condition. The review also summarises the main benefits, challenges, and limitations of such systems. It is clear from the results that AI as a component of precision farming has the potential to revolutionise practices and non-invasive methods of monitoring. However, continued research and development will help to address the limitations and challenges, including the need for data management and standardisation. By addressing the problems associated with image-based methodologies such as poor video quality and a lack of behavioural and disease monitoring trials using AI in industry, fish welfare could be greatly improved. Chapter 2 tests two image-based methodologies for estimating body mass index (BMI) of lumpfish and ballan wrasse. The first was ImageJ, a software which can be used to measure fish from photographs that include a scalebar. The results from this study showed that measurements obtained from ImageJ could be used to estimate BMI with good accuracy. The second was a new Artificial Intelligence (AI) system that was developed by Visifish, which determined fish measurements from videos. Visifish are a bioinformatics company that use machine vision and deep learning techniques to assess population, growth and behaviour. The AI was only able to provide measurements for 21 out of the 120 fish that were filmed. The results showed that the AI was unable to provide measurements that were accurate enough to determine true BMI. This was likely due to the limited training data set. To validate the AI, more work will need to be carried out to developit further, including larger sample sizes and more videos. Chapter 3 involves 2 experiments with the goal of determining whether ‘personality components’ can be identified and potentially used for broodstock selection for the use in the aquaculture industry. The first experiment involved recording individuals of both species and their behavioural responses to different stimuli. Their responses were used to label them with personality components. The second experiment involved recording the same fish’ response to artificial Atlantic salmon models with artificial sea lice attached on to their flank. The personality components were then compared with the responses to the salmon models to determine whether the personality components could relate to good delousing behaviours. The results showed that lumpfish were significantly bolder, more aggressive, more social, and less anxious than the ballanwrasse. The results also showed that bold, social, and not anxious individuals had more interactions with the salmon model carrying a large number of sea lice. This suggests that these personality components could be selected and used to breed potentially good delousing behaviours. Chapter 4 involves using a shuttle-box and machine vision system to determine the temperature preference and thermal niche of ballan wrasse. The temperature preference and thermal niche of lumpfish was not explored in this thesis as it has already been studied using the same system and is awaiting publishing. The results concluded that the preferred thermal niche for ballan wrasse is between 12.1°C and 13.8°C. The optimal temperature preference was 12.8°C. These results could be used to aid decisions about deployment times and locations for ballan wrasse in the aquaculture industry.</abstract><type>E-Thesis</type><journal/><volume/><journalNumber/><paginationStart/><paginationEnd/><publisher/><placeOfPublication>Swansea University, Wales, UK</placeOfPublication><isbnPrint/><isbnElectronic/><issnPrint/><issnElectronic/><keywords>Aquaculture, cleaner fish, AI, machine vision, behavioural indicators</keywords><publishedDay>21</publishedDay><publishedMonth>3</publishedMonth><publishedYear>2024</publishedYear><publishedDate>2024-03-21</publishedDate><doi/><url/><notes>A selection of content is redacted or is partially redacted from this thesis to protect sensitive and personal information.</notes><college>COLLEGE NANME</college><CollegeCode>COLLEGE CODE</CollegeCode><institution>Swansea University</institution><supervisor>De Leaniz, C. G., and De Olmo, S. C.</supervisor><degreelevel>Master of Research</degreelevel><degreename>MSc by Research</degreename><degreesponsorsfunders>KESS 2, Visifish</degreesponsorsfunders><apcterm/><funders>KESS 2, Visifish</funders><projectreference/><lastEdited>2024-06-20T14:32:23.7906497</lastEdited><Created>2024-04-25T15:16:44.2674160</Created><path><level id="1">Faculty of Science and Engineering</level><level id="2">School of Biosciences, Geography and Physics - Biosciences</level></path><authors><author><firstname>ISLA</firstname><surname>MONAGHAN</surname><order>1</order></author></authors><documents><document><filename>Under embargo</filename><originalFilename>Under embargo</originalFilename><uploaded>2024-06-20T14:29:54.8656265</uploaded><type>Output</type><contentLength>5428523</contentLength><contentType>application/pdf</contentType><version>E-Thesis</version><cronfaStatus>true</cronfaStatus><embargoDate>2025-05-21T00:00:00.0000000</embargoDate><documentNotes>Copyright: The Author, 2024, Isla Monaghan.</documentNotes><copyrightCorrect>true</copyrightCorrect><language>eng</language></document></documents><OutputDurs/></rfc1807> |
| spelling |
v2 66199 2024-04-25 Using machine vision and behavioural indicators to monitor cleaner fish welfare in the Atlantic salmon (Salmo salar) aquaculture industry fd52564ded9420c92e1edcb4d81ff39d ISLA MONAGHAN ISLA MONAGHAN true false 2024-04-25 The aim of this thesis was to explore how machine vision and behaviour indicators could be used to monitor welfare of Ballan wrasse (Labrus bergyltus) and Lumpfish (Cyclopterus lumpus), which are used as cleaner fish in the Atlantic salmon industry. This thesis is divided into 4 chapters. Chapter 1 is a systematic literature review on the use of artificial intelligence (AI) in the aquaculture industry. The review looks into examples of different uses for AI in the industry, including intelligent feeding, monitoring of water quality, behaviour and disease, and assessment of body condition. The review also summarises the main benefits, challenges, and limitations of such systems. It is clear from the results that AI as a component of precision farming has the potential to revolutionise practices and non-invasive methods of monitoring. However, continued research and development will help to address the limitations and challenges, including the need for data management and standardisation. By addressing the problems associated with image-based methodologies such as poor video quality and a lack of behavioural and disease monitoring trials using AI in industry, fish welfare could be greatly improved. Chapter 2 tests two image-based methodologies for estimating body mass index (BMI) of lumpfish and ballan wrasse. The first was ImageJ, a software which can be used to measure fish from photographs that include a scalebar. The results from this study showed that measurements obtained from ImageJ could be used to estimate BMI with good accuracy. The second was a new Artificial Intelligence (AI) system that was developed by Visifish, which determined fish measurements from videos. Visifish are a bioinformatics company that use machine vision and deep learning techniques to assess population, growth and behaviour. The AI was only able to provide measurements for 21 out of the 120 fish that were filmed. The results showed that the AI was unable to provide measurements that were accurate enough to determine true BMI. This was likely due to the limited training data set. To validate the AI, more work will need to be carried out to developit further, including larger sample sizes and more videos. Chapter 3 involves 2 experiments with the goal of determining whether ‘personality components’ can be identified and potentially used for broodstock selection for the use in the aquaculture industry. The first experiment involved recording individuals of both species and their behavioural responses to different stimuli. Their responses were used to label them with personality components. The second experiment involved recording the same fish’ response to artificial Atlantic salmon models with artificial sea lice attached on to their flank. The personality components were then compared with the responses to the salmon models to determine whether the personality components could relate to good delousing behaviours. The results showed that lumpfish were significantly bolder, more aggressive, more social, and less anxious than the ballanwrasse. The results also showed that bold, social, and not anxious individuals had more interactions with the salmon model carrying a large number of sea lice. This suggests that these personality components could be selected and used to breed potentially good delousing behaviours. Chapter 4 involves using a shuttle-box and machine vision system to determine the temperature preference and thermal niche of ballan wrasse. The temperature preference and thermal niche of lumpfish was not explored in this thesis as it has already been studied using the same system and is awaiting publishing. The results concluded that the preferred thermal niche for ballan wrasse is between 12.1°C and 13.8°C. The optimal temperature preference was 12.8°C. These results could be used to aid decisions about deployment times and locations for ballan wrasse in the aquaculture industry. E-Thesis Swansea University, Wales, UK Aquaculture, cleaner fish, AI, machine vision, behavioural indicators 21 3 2024 2024-03-21 A selection of content is redacted or is partially redacted from this thesis to protect sensitive and personal information. COLLEGE NANME COLLEGE CODE Swansea University De Leaniz, C. G., and De Olmo, S. C. Master of Research MSc by Research KESS 2, Visifish KESS 2, Visifish 2024-06-20T14:32:23.7906497 2024-04-25T15:16:44.2674160 Faculty of Science and Engineering School of Biosciences, Geography and Physics - Biosciences ISLA MONAGHAN 1 Under embargo Under embargo 2024-06-20T14:29:54.8656265 Output 5428523 application/pdf E-Thesis true 2025-05-21T00:00:00.0000000 Copyright: The Author, 2024, Isla Monaghan. true eng |
| title |
Using machine vision and behavioural indicators to monitor cleaner fish welfare in the Atlantic salmon (Salmo salar) aquaculture industry |
| spellingShingle |
Using machine vision and behavioural indicators to monitor cleaner fish welfare in the Atlantic salmon (Salmo salar) aquaculture industry ISLA MONAGHAN |
| title_short |
Using machine vision and behavioural indicators to monitor cleaner fish welfare in the Atlantic salmon (Salmo salar) aquaculture industry |
| title_full |
Using machine vision and behavioural indicators to monitor cleaner fish welfare in the Atlantic salmon (Salmo salar) aquaculture industry |
| title_fullStr |
Using machine vision and behavioural indicators to monitor cleaner fish welfare in the Atlantic salmon (Salmo salar) aquaculture industry |
| title_full_unstemmed |
Using machine vision and behavioural indicators to monitor cleaner fish welfare in the Atlantic salmon (Salmo salar) aquaculture industry |
| title_sort |
Using machine vision and behavioural indicators to monitor cleaner fish welfare in the Atlantic salmon (Salmo salar) aquaculture industry |
| author_id_str_mv |
fd52564ded9420c92e1edcb4d81ff39d |
| author_id_fullname_str_mv |
fd52564ded9420c92e1edcb4d81ff39d_***_ISLA MONAGHAN |
| author |
ISLA MONAGHAN |
| author2 |
ISLA MONAGHAN |
| format |
E-Thesis |
| publishDate |
2024 |
| institution |
Swansea University |
| college_str |
Faculty of Science and Engineering |
| hierarchytype |
|
| hierarchy_top_id |
facultyofscienceandengineering |
| hierarchy_top_title |
Faculty of Science and Engineering |
| hierarchy_parent_id |
facultyofscienceandengineering |
| hierarchy_parent_title |
Faculty of Science and Engineering |
| department_str |
School of Biosciences, Geography and Physics - Biosciences{{{_:::_}}}Faculty of Science and Engineering{{{_:::_}}}School of Biosciences, Geography and Physics - Biosciences |
| document_store_str |
0 |
| active_str |
0 |
| description |
The aim of this thesis was to explore how machine vision and behaviour indicators could be used to monitor welfare of Ballan wrasse (Labrus bergyltus) and Lumpfish (Cyclopterus lumpus), which are used as cleaner fish in the Atlantic salmon industry. This thesis is divided into 4 chapters. Chapter 1 is a systematic literature review on the use of artificial intelligence (AI) in the aquaculture industry. The review looks into examples of different uses for AI in the industry, including intelligent feeding, monitoring of water quality, behaviour and disease, and assessment of body condition. The review also summarises the main benefits, challenges, and limitations of such systems. It is clear from the results that AI as a component of precision farming has the potential to revolutionise practices and non-invasive methods of monitoring. However, continued research and development will help to address the limitations and challenges, including the need for data management and standardisation. By addressing the problems associated with image-based methodologies such as poor video quality and a lack of behavioural and disease monitoring trials using AI in industry, fish welfare could be greatly improved. Chapter 2 tests two image-based methodologies for estimating body mass index (BMI) of lumpfish and ballan wrasse. The first was ImageJ, a software which can be used to measure fish from photographs that include a scalebar. The results from this study showed that measurements obtained from ImageJ could be used to estimate BMI with good accuracy. The second was a new Artificial Intelligence (AI) system that was developed by Visifish, which determined fish measurements from videos. Visifish are a bioinformatics company that use machine vision and deep learning techniques to assess population, growth and behaviour. The AI was only able to provide measurements for 21 out of the 120 fish that were filmed. The results showed that the AI was unable to provide measurements that were accurate enough to determine true BMI. This was likely due to the limited training data set. To validate the AI, more work will need to be carried out to developit further, including larger sample sizes and more videos. Chapter 3 involves 2 experiments with the goal of determining whether ‘personality components’ can be identified and potentially used for broodstock selection for the use in the aquaculture industry. The first experiment involved recording individuals of both species and their behavioural responses to different stimuli. Their responses were used to label them with personality components. The second experiment involved recording the same fish’ response to artificial Atlantic salmon models with artificial sea lice attached on to their flank. The personality components were then compared with the responses to the salmon models to determine whether the personality components could relate to good delousing behaviours. The results showed that lumpfish were significantly bolder, more aggressive, more social, and less anxious than the ballanwrasse. The results also showed that bold, social, and not anxious individuals had more interactions with the salmon model carrying a large number of sea lice. This suggests that these personality components could be selected and used to breed potentially good delousing behaviours. Chapter 4 involves using a shuttle-box and machine vision system to determine the temperature preference and thermal niche of ballan wrasse. The temperature preference and thermal niche of lumpfish was not explored in this thesis as it has already been studied using the same system and is awaiting publishing. The results concluded that the preferred thermal niche for ballan wrasse is between 12.1°C and 13.8°C. The optimal temperature preference was 12.8°C. These results could be used to aid decisions about deployment times and locations for ballan wrasse in the aquaculture industry. |
| published_date |
2024-03-21T14:32:23Z |
| _version_ |
1802387160796495872 |
| score |
11.035874 |