ANIHWAcall3. Validation of automated welfare assessment for poultry

Lead Research Organisation: University of Oxford
Department Name: Zoology

Abstract

The welfare of food animals is a high priority for European consumers who increasingly see high welfare as an essential part of sustainable, healthy food production. Animal welfare can potentially be greatly enhanced by using technology capable of giving continuous information about the health and welfare of the animals concerned, but so far poultry production has lagged behind the rest of the agricultural sector and still lacks widely available automated ways of assessing animal welfare. In commercial broiler (meat) chicken houses, group sizes are so large (flocks of up to 50,000 are common in Europe) and the margins on each bird so low that it is not feasible to assess the welfare of each animal individually as is possible with cows or pigs, so a different approach to automated welfare assessment is needed.
One promising method of automatically assessing the welfare of poultry that has already achieved initial validation on commercial farms involves using the pattern of 'optical flow' created by the natural movements of chicken flocks as they move around house. The initial validation has shown that optical flow can predict key welfare outcomes in broiler chickens days and even weeks in advance of other types of measurements. It can even distinguish which flocks are at risk from subsequently testing positive for Campylobacter. The system could potentially have worldwide impact if taken up and used as a standard welfare assessment tool for poultry. However, if the system is to realise its full potential for measuring and improving poetry welfare, it needs to be subject to a second stage of validation. That is the aim of this project.
The project will do this (i) by testing the optical flow system on a wide range of farms and conditions across Europe (Switzerland, France and the UK) and then (ii) by comparing it with other already recognised but labour-intensive methods for assessing welfare at flock level, such as 'gait scoring'. (iii) An important component of the project will be to relating the welfare of birds measured at the individual level to optical flow and other measures at flock level. Since 'good' or 'bad' welfare is experienced by individuals not groups we need to establish how well optical flow reflects individual bird welfare. This will be done by applying a range of behavioural tests to individual birds, looking at their physical health and then seeing how well these are correlated with whole flock behaviour. (iv) We will relate optical flow to infection by key diseases such as Campylobacter and Salmonella, both of which impact gut health and are of great concern to the poultry industry. (v) Finally,using the information collected, a database will be set up so that optical flow data can be seen and evaluated in the context of production, health, management and other flock attributes with the aim of establishing 'norms' for good flock welfare.
The project consortium consists of teams from the UK, France, Switzerland and the Czech Republic and includes experts on behaviour, welfare, stress physiology, poultry health. immunology as well as specialists in Salmonella infection, genotyping of Campylobacter and DNA technology for analysing gut microbiota and health. Although diverse in their individual interests, the members of the consortium are united in their determination to improve poetry health and welfare by careful and validated use of technology.

Technical Summary

The technical objectives of the project as a whole are set out in the full ANIHWA proposal (attached)
The UK contribution will be

1. To collect optical flow, health, welfare, management and production data, including samples for microbiological analysis, from commercial broiler farms in the UK and to be responsible for running equipment on Swiss and French, 25 farms. Optical flow data will be obtained via an existing camera/computer system already developed in Oxford. Behaviour, health and gait scores will be collected in a standard form. Data on water use, % mortality weight gain, and production data specially hockburn and pododermatitis will be obtained from the production companies.

2. To construct a database of all data collected from different farms across Europe with a view to developing flock 'norms' for healthy high welfare flocks and criteria for flocks that depart from this norm and to being able to relate welfare and disease to gut health. It will include existing measures used by producers for flock management such as water use and growth rate. the aim will be to develop means of identifying 'norms' for against which good welfare can be assessed.

3. Individual partners will be responsible for the initial microbiological screening of the samples they collect. Sub-sampls will be selected for further molecular analysis such as qPCR,analysis of genetic diversity at a single hypervariable focus (e.g. Por A gene in Campylobacter) and 16S microbiome analysis of all species in the gut.

Planned Impact

The aim of the project is to produce a flock management tool that will be of direct use to farmers, veterinarians and researchers. It will enable producer companies to set standards of health and welfare within their companies and to put 'early warning' systems in place for their flocks. These will in turn allow them to demonstrate to their customers (such as supermarkets, food outlets and the general public) that health and welfare are priorities for them in poultry production. The benefits to the end users (farmers) will be partly directly financial (better flock management leading to lower mortality, less waster and more efficient production) and party due to more intangible reputational advantages such as the public image benefits of reducing antibiotics and being seen to achieve higher standards of animal welfare.
Since much of the work will be carried out on commercial farms in France, Switzerland and the UK, the work will involve close collaboration with commercial broiler producers who will thus learn from direct experience what the system is capable of doing and what benefits it brings to their own workforce.This will be our 'front line publicity and, initially, our main way of having an impact on end-users. the highly structured nature of the poetry industry (only a relatively small number of companies are responsible for a large proportion of chicken production) means that dissemination of the results through the industry can be quickly accomplished through contact with a few key players. A key part of our impact plan is to keep the companies we work with fully informed and up to date with what we are doing as they are the ones we need to influence.
To widen the impact of our work still further, we will attend international trade fairs which enable us to make contact with commercial representatives across Europe and the wider world. For dissemination to more academic audiences, we will attend national and international conferences on behaviour, poetry health, animal welfare band environmental microbiology.
As the work has broad implications in multiple areas and will interest researchers in diverse fields. We will therefore publish the results in a wide range of journals including journals on applied behaviour, foodbourne human diseases and gut microbiology. In view of the public interest in improving animal welfare, we will also publish in journals that are accessible to the genial public, to legislators and policy makers.

Publications

10 25 50

publication icon
Dawkins (2019) Animal Welfare as Preventative Medicine in Animal Welfare

publication icon
Dawkins (2019) Animal Welfare as Preventative Medicine in Animal Welfare

publication icon
Dawkins M (2017) Animal welfare with and without consciousness in Journal of Zoology

publication icon
Dawkins M (2021) Optical flow, behaviour and broiler chicken welfare in the UK and Switzerland in Applied Animal Behaviour Science

publication icon
Fiddaman S (2022) Adaptation and Cryptic Pseudogenization in Penguin Toll-Like Receptors in Molecular Biology and Evolution

publication icon
Ryan H (2022) Reconstructing the history of helminth prevalence in the UK in PLOS Neglected Tropical Diseases

 
Description Key findings:
1. We have been able to replicate our previous results relating optical flow to welfare outcomes in broiler chickens in both the Uk and the very different farm conditions found in Switzerland.This was an important step in validating OpticFlock as a means of assessing broiler welfare. It shows that it is robust and applicable in a variety of farm conditions
2. We have shown that optical flow patterns of flocks are directly correlated with the behaviour of individual birds. A sample fo 16 birds from each flock were put through a series of individual behaviour tests. We found that birds that performing slowly in behaviour tests (moving down a runway and standing in water) came from flocks with low mean and high kurtosis of optical flow whereas birds that moved fast can from flocks with high mean and low kurtosis.
3. Using deep sequencing genetic techniques, we have shown that all broiler flocks tested carried Campylobacter even when under 8 days of age, , but in such minute quantities that they cannot be detected by standard culture methods. Only some flocks went on to later develop 'florid' growth of Campylobacter that could be detected by these future methods.
Exploitation Route The technology is already being taken forward. We signed a Development Licence with the Munters Corporation which is now in process of commercialising OpticFock. . Munters is a global company with marketing outlets throughout the world so this would be a route by which the technology could be widely used. We have also applied the same approach to the problem of predicting tail-biting in pigs
Sectors Agriculture, Food and Drink,Digital/Communication/Information Technologies (including Software),Environment

URL http://users.ox.ac.uk/~snikwad/index.html
 
Description It has lead to a collaboration with the Munters Corporation to develop OpticFlock commercially. Our research has also been supported by a grant from the Foundation for Food and Agricultural Research in collaboration with Tyson Foods
Sector Agriculture, Food and Drink,Environment,Healthcare
Impact Types Societal,Economic

 
Description BBSRC IAA University of Oxford
Amount £300,000 (GBP)
Funding ID BB/S50676X/1 
Organisation Biotechnology and Biological Sciences Research Council (BBSRC) 
Sector Public
Country United Kingdom
Start 03/2018 
End 03/2021
 
Description EPSRC Impact Acceleration Award
Amount £89,396 (GBP)
Funding ID EP/R511742/1 
Organisation University of Oxford 
Sector Academic/University
Country United Kingdom
Start 07/2018 
End 07/2019
 
Description FFAR SMART Broiler
Amount $232,064 (USD)
Funding ID Sb-0000000005 
Organisation Foundation for Food and Agriculture Research 
Sector Charity/Non Profit
Country United States
Start 01/2020 
End 06/2021
 
Description OpticFlock: de-risking for market evaluation
Amount £89,396 (GBP)
Funding ID EP/R511742/1 
Organisation University of Oxford 
Department EPSRC Impact Acceleration Account
Sector Academic/University
Country United Kingdom
Start 07/2018 
End 07/2019
 
Title OpticFlock: automated assessment of broiler chicken welfare 
Description OpticFlock is a camera/computer system that can be placed inside commercial chicken houses and gives a continuous 'verdict' on the welfare of the birds throughout their lives. It works by taking visual images from a cctv camera and running them through an 'optical flow' algorithm pre-stalled on a small computer. 'Optical flow' measures the rate of change of image brightness and is able to pick up the statistical patterns of movement made by flocks as they move. Our background research has shown that certain key optical flow statistics- particularly the mean rate of movement and the kurtosis (a measure of variation) are highly predictive of key welfare outcomes (especially hockburn and lameness) and are even able to predict which flocks are at greatest risk of later testing positive for Campylobacter, days or weeks before this is detectable by standard culture methods. 
Type Of Material Physiological assessment or outcome measure 
Year Produced 2012 
Provided To Others? Yes  
Impact Negotiations for a Development Licence with the Munters Corporation for the commercialization of OpticFlock 
 
Title welfare assessment tool (OPTICFLOCK) 
Description Use of camera technology to assess welfare in broiler chickens. We have developed a system (OPTICFLOCK) that analyses video data on farms and send the results to a web portal. Farmers can be sent a daily updates the state of their flocks in comparison to standard or reference flocks. 
Type Of Material Technology assay or reagent 
Year Produced 2015 
Provided To Others? Yes  
Impact We have received a great deal of interest from the commercial world and have entered into development agreements with some multi-national companies. We are still validating the application to a new species (pigs) 
URL http://opticflock.com
 
Description Commercial development of OpticFlock 
Organisation Munters Ltd
Country United Kingdom 
Sector Private 
PI Contribution We have shared our data on optical flow with Munters so that we could all compare our results on welfare with their results on environmental variables. A year-long trial was conducted to allow sufficient data to be collected for such a comparison
Collaborator Contribution As a result of this preliminary collaboration, Munters indicted that they would be interested in proceeding to discussion over a Development Licence in which they would look at the feasibility of incorporating OpticFlock data onto their Sonar/Echo-Mesh platform. we are in process of negotiating the development licence, which could immensely befeit the commercialisation of OpticFlock.
Impact This collaboration is multidisciplinary: Animal welfare, software development, statistical analysis The outputs are currently being developed. This week (February 2020) we have just received the equipment from Munters that will enable us to see whether it can receive and display OpticFlock data. Tests on a trial farm in the US are planned for later this year.
Start Year 2017
 
Description Commercial test of OPTICFLOCK during feed trials 
Organisation DuPont
Country United States 
Sector Private 
PI Contribution We have agreed to analyse cctv footage for trials due to start at the beginning of March 2019
Collaborator Contribution Dupongt are setting up the trials, making the video recordings and suppling them to us
Impact Trials pallned to start in March 2019 so now outputs yet
Start Year 2018
 
Description Development of camera/OPTICFLOCK system 
Organisation Agri-cctv Ltd
Country United Kingdom 
Sector Private 
PI Contribution We supply technical advice and analysis
Collaborator Contribution Agri-cctv install cameras and our equipment in chicken shed
Impact Trials started in May 2017 and finished in December 2018. The outcome was considerable interest fro a major international compaby, with whom a licensing arrangement is currently being negotiated
Start Year 2017
 
Description In-kind help with broiler welfare trials 
Organisation Cargill
Country United States 
Sector Private 
PI Contribution We have been helping to develop automated ways of measuring welfare
Collaborator Contribution The company has provided access to broiler houses and help from farm staff They also provide us with faecal samples for gut analysis
Impact Useful results on behaviour and chicken disease and welfare
Start Year 2012
 
Description Integration of OPTICFLOCK data onto a commercially available dashbiard 
Organisation Munters Ltd
Country United Kingdom 
Sector Private 
PI Contribution Knowledge of flock behaviour
Collaborator Contribution We are negotaiation what they they will provide but we expect it to be help with development
Impact Please note that we are collaborating with Munters internationally but not Munters UK specifically (but this is the only option yoyr system allows!). We are currently negotiating a lience agreement for development
Start Year 2018
 
Description On-farm trial of OpticFlock 
Organisation Avara Foods Ltd
Country United Kingdom 
Sector Private 
PI Contribution We provided them with information about the welfare of their flocks and the presence (and genotypes) of Campylobacter with the aim of helping them to trace the source of Campylobacter infection. This has resulted the important result that very young birds in all flocks tested appear to carry Campylobacter but in such minute quantities that these are undetectable by normal culture methods.
Collaborator Contribution Permitting us to put cameras in their chicken houses, providing end-of-flock production data on each flock and collecting faecal samples for us.
Impact The collaboration is multidisciplinary: Animal welfare (Department of Zoology), Microbiology (Department of Zoology), software (Dept. Engineering Science). statistics )DEparmemt of Statistics. The demonstration that Campylobacter is present in minute quantities even in chicks as young as 3 days hold is potentially of great importance in developing control strategies. We are currently in discussion with Cargill (now AVARA) about further testing of the hypothesis that Campylobacter control may be best achieved through improved welfare.
Start Year 2016
 
Description On-farm trial of OpticFlock 
Organisation Tyson Foods
Country United States 
Sector Private 
PI Contribution We provided them with information about the welfare of their flocks and the presence (and genotypes) of Campylobacter with the aim of helping them to trace the source of Campylobacter infection. This has resulted the important result that very young birds in all flocks tested appear to carry Campylobacter but in such minute quantities that these are undetectable by normal culture methods.
Collaborator Contribution Permitting us to put cameras in their chicken houses, providing end-of-flock production data on each flock and collecting faecal samples for us.
Impact The collaboration is multidisciplinary: Animal welfare (Department of Zoology), Microbiology (Department of Zoology), software (Dept. Engineering Science). statistics )DEparmemt of Statistics. The demonstration that Campylobacter is present in minute quantities even in chicks as young as 3 days hold is potentially of great importance in developing control strategies. We are currently in discussion with Cargill (now AVARA) about further testing of the hypothesis that Campylobacter control may be best achieved through improved welfare.
Start Year 2016
 
Description Research trials for commercialising OPTICFLOCK 
Organisation Tuffigo-Rapidex
Country France 
Sector Private 
PI Contribution We set up OPTICFLOCK on five French farms and planned to supply Tuffigo with optical flow data. Unfortunately there were problem video recordings, which turned out to be unsuitable for optical flow analysis. We believe that the NAS drives storing the data were repeatedly switched on and off and the computers were therefore incorrectly shut-down. After repeated attempts to retrieve data from corrupted NAS drives. The trials were abandoned.
Collaborator Contribution Tuffigo-rapidex were to have supply us with house data an production data on the same flocks
Impact Data was collected Nov.2016-August 2017 but there were technical problems that prevented us collecting as good a dataset as we would have liked. Other data collected (e.g. Campylobacter) will be used in subsequent publications.
Start Year 2016
 
Description Research trials in Arkansas 
Organisation Tyson Foods
Country United States 
Sector Private 
PI Contribution We set up a research trial at the Tyson Research Farm in Arkansas to see whether OpticFlock would work on US broiler farms. We discussed with Tyson the best way to measure animal welffare
Collaborator Contribution Tyson supplied technical help in setting up our equipment and also supplied us with a lot of important information about the chicken flocks we were recording from . Due to Covid, we could not visit the farm and had to rely entirely on the Tyson staff for installation of equipment and for sending us data. Their support was invaluable.
Impact We were able to show that OpticFlock worked successfully on a US broiler farm, despite differences in breed and management system. This was an important step in its commercial development. As a result of this, we were invited to apply for further funding from FFAR (Foundation for Food and Agricultural Research) under their SMART Broiler program, which is co-sponosred by McDonalds. This will enable us to carry on research in the US, UK and Hungary
Start Year 2019
 
Description Research trials on commercialising OPTICFLOCK 
Organisation SKOV A/S
Country Denmark 
Sector Private 
PI Contribution We have downloaded data from SKOV and have analysed it
Collaborator Contribution SKOV have supplied us with video records and production data for commercial broiler flocks
Impact SKOV want to enter into a licence agreement but we are still negotiating this. They wanted us to pay fro any further developments so we do not think this is a favoruable route to go down.
Start Year 2016
 
Title OpticFlock 
Description OpticFlock is a camera/computer system for automatic assessment for broiler chicken welfare that delivers a welfare 'verdict' throughout the life of a flock. It takes moving images from a cctv camera and analyzes the statistical patterns of movement made by the flocks as they move. The data from a given flock are compared with those of reference flocks with so that a farmer can see whether the flock is behaving as normal for high welfare flocks or whether there are problems. This early warning of welfare/disease issues can be used to intervene to prevent problems becoming more serious. 
IP Reference  
Protection Trade Mark
Year Protection Granted 2016
Licensed Commercial In Confidence
Impact We are in process of negotiating a Development Licence with a view to trials in the US this year. By December, progress will be reviewed to see whether both parties wish to proceed to a full licensing agreement.
 
Title OPTICFLOCK 
Description We have developed OPTICFLOCK software so that it automatically analyses the moments of chicken flocks, analyses that moment on farms and reports the results to a web portal, which in turn conveys updated information to a farmer on a daily basis. We have applied for trademark registration for OPTICFLOCK. 
Type Of Technology Webtool/Application 
Year Produced 2017 
Impact We are just about (March 2017) to test the first prototype, 
URL http://opticflock.com
 
Description Automated assessment of welfare 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Industry/Business
Results and Impact Discussion with British Egg Industry. London. 28 November 2018
Year(s) Of Engagement Activity 2018
 
Description Conference for poultry industry 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Industry/Business
Results and Impact This was a well attended meeting of the Poultry Science Association in Orlando, Florida, USA at which there was an audience from many different sectors and with many representatives from the poultry industry. There was a special section on poultry welfare.
Year(s) Of Engagement Activity 2017
 
Description How animal behaviour can help us assess the health and well-being of poultry 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Professional Practitioners
Results and Impact invited plenary talk to Poultry Health and Management Conference, Loughborough, UK. 20th Novermber, 2918.
Year(s) Of Engagement Activity 2018
 
Description Innovative system for monitoring poultry health 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Professional Practitioners
Results and Impact Invited plenary talk to the Italian Branch of the World Poultry Science Association in Perugia, Italy. 6th April, 2018.
Year(s) Of Engagement Activity 2018
 
Description Keynote talk to World Animal Protection conference, Bangkok, Thailand 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Third sector organisations
Results and Impact Keynote speech on assessing animal welfare to the World Animal Protection (formerly WSPA) aimed at facilitating links to industry
Year(s) Of Engagement Activity 2016
 
Description Presentation to Perdue Farms, USA 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Industry/Business
Results and Impact Perdue Farms organised a 'poultry summit' and open day and invited me to be a keynote speaker
Year(s) Of Engagement Activity 2017
 
Description Talk to Norfolk poultry farmers 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach Regional
Primary Audience Public/other audiences
Results and Impact Talk to Norfolk poultry farmers
Year(s) Of Engagement Activity 2016
 
Description Talk to international group at London Airport 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Professional Practitioners
Results and Impact Small international conference
Year(s) Of Engagement Activity 2016
 
Description WPVA Plenary talk 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Professional Practitioners
Results and Impact Large International Conference of the World Poultry veterinary Association in Edinburgh 3-6 September 2017
Year(s) Of Engagement Activity 2017
 
Description automated assessment of poultry welfare 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Professional Practitioners
Results and Impact Invited talk to the Animal Welfare Research Network conference in Birmingham. 3rd September, 2018
Year(s) Of Engagement Activity 2018