Defra - Department for Environment, Food and Rural Affairs.

Science Search

Science and Research Projects

Return to Science Search homepage   Return to Project List

Feasibility study to improve badger monitoring through automatic recognition - SE3295

Description
Bovine tuberculosis (bTB) is described by government as the most pressing animal health problem in the UK. Dealing with bTB in England costs taxpayers over £100 million a year and required the culling of 28,000 cattle in 2015. The disease and the measures used to control it, can have considerable social and economic consequences for farmers and impacts on the economic security of the dairy and beef sectors. bTB is principally a disease of cattle, but there are several places worldwide where free-ranging wildlife are reservoirs of infection, namely brushtail possums in New Zealand, wood bison and elk in Canada, African buffalo in South Africa, white tailed deer in the United States and European badgers (Meles meles) in the United Kingdom and Republic of Ireland.
The Strategy to Achieve Officially TB Free Status (OTF) for England was published in 2014 and established a framework for controlling all routes of transmission of the disease. These are principally cattle-cattle, but also include cattle-badger, badger-badger and badger-cattle. In addition to countrywide surveillance of the cattle population, monitoring of the badger population plays a key role in helping to understand disease transmission routes and identify mechanisms to prevent opportunities for disease transfer between species. There remains a lack of clear understanding of how disease is transmitted between cattle and badgers and vice versa (Godfay et al 2013), which makes the development of new methods to break the cycle of infection more challenging.
Previous research has shown that badgers may visit farm buildings and can come into direct contact with cattle, although the extent to which this can occur varies geographically. Research also indicates that farmers tended to underestimate the level and frequency of badger visits to their farm holdings, suggesting a lack of awareness of the need to prevent badger access to buildings and feed stores.
Much of the badger monitoring research on farms and at research facilities (such as Woodchester Park) has relied on CCTV and camera-traps to capture badger activity. The camera may be triggered by humans, cattle or wildlife other than badgers, so this approach can produce large number of images that need to be reviewed by staff. Also, camera-trap technology does not give the wider context on badger behaviour (e.g. where badgers access and visit on farm) and due to the large number of images that would be generated, researchers tend not to place cameras in areas with high cattle activity, which prevents efficient analysis of cattle-badger interactions.
To address these challenges, this project proposes to pilot the use of machine learning technology to develop cheaper, more effective badger recognition on live-feed cameras. The ability to specifically identify badgers would reduce the amount of time and effort required for monitoring and allow cameras to be placed in areas of high activity. Automated detection, using cheaper cameras, would increase the number of farms and locations that can be monitored and could lead to further developments such as a hand-held app, alerting farmers to badger activity on farm. This would benefit disease control strategies by providing more efficient badger monitoring on farms to help farmers identify areas for intervention (e.g. to prevent badger entry or badger-cattle contact), thus reducing opportunities for disease transmission.
Objective
Aim:
To develop and pilot automatic recognition software capable of identifying and isolating badger activity

Objectives:
1. Design and deliver an automated badger detection camera capable of improving badger monitoring activities
2. Investigate opportunities to further exploit the approach to improve badger monitoring, including the potential to develop novel on-farm applications to improve farm biosecurity
Time-Scale and Cost
From: 2017

To: 2018

Cost: £10,573
Contractor / Funded Organisations
University of Sheffield
Keywords
Animals              
Fields of Study
Animal Health