Because wild boars are often involved in road accidents and cause damage to agricultural crops, their presence in the province of Overijssel is closely monitored. However, the location of the animals, who hide during the day, appears to be difficult to map. DesignLab students now have a made a first step towards spotting the population using drones.
In this project, DesignLab University of Twente worked together with Space53 and the Province of Overijssel. The Province supplied knowledge on the habitat of wild boars, while Drone Flight Company, as a participant in the Space53 drone innovation cluster, made two drone flights above a nature area inhabited by boars. Under the supervision of DesignLab, four UT students developed an approach to recognise wild boars automatically using the images produced by drones. Ecologist-geologist Dr Panagiotis Nyktas of the faculty of Geo-Information Science and Earth Observation (ITC) provided the students with advice.
Active at night
When developing their method, the students encountered various challenges. As wild boars hide under the vegetation during the day, they are hardly visible – or even invisible – to drones. One of the results of the study is the advice to make thermal images during winter nights. Wild boars are active at night. During cold nights without foliage to hide under, they are better recognisable on thermal images.
Flying with an octocopter
In addition, the student team proposes to use machine learning to achieve automatic recognition of boars. This, however, requires images of the animals, something that could not be obtained during this study. This is why the students advise to use a specific type of drone: the octocopter, which offers sufficient solidity, control, and coverage. Stability is very important for spotting wild boars, especially when the drone is flying slowly to get a better picture of individual animals.
Approach is feasible
The conclusion of the study is that spotting wild boards using drones and machine learning is certainly feasible. It does involve quite a number of challenges, because of the behaviour of the animals. Once visual material of boars is available, it will be possible to follow the approach proposed by the students.
During a follow-up phase, the partners will explore the proposed approach and new drone flights will be made. DesignLab prepares a follow-up using Citizen Science for wild monitoring, so interested citizens can help map out the population.
DesignLab’s contribution to Mapping Overijssel
This project is part of Mapping Overijssel, a joint initiative of the University of Twente and the Innovation and Data Lab (ID-Lab) of the Province of Overijssel. The ID-Lab aims to make a contribution to a smarter implementation of the social challenges of the province by conducting experiments with data and data technology.
In a series of experiments in the area of regional digitisation, a new approach was tested by bringing research and policy closer together by designing and working in a multidisciplinary environment. DesignLab contributes to this project by providing scientific, technological and creative insights. In this way, we realise influential solutions for social issues.
For more information about the project, please contact project manager Maya van den Berg.