Since 2013, Wageningen University has been conducting research into the behavior of wild animals in De Hoge Veluwe National Park using camera traps. It soon became apparent that the cameras provided such a large amount of images (more than 1 million per year) that it was impossible for a limited number of researchers and volunteers to view and review them. It was then decided to enlist the help of interested citizens.
On May 29, 2018, De Hoge Veluwe National Park started Snapshot Hoge Veluwe together with Wageningen University. Initially, the photos could only be viewed via the Park's website, but a year later they were put online by Zooniverse, the largest and most popular platform for citizen science in the world. Snapshot Hoge Veluwe was the first European National Park to be located on Zooniverse.
More than fifteen thousand registered (and a much larger number of unregistered) citizen scientists from the Netherlands and abroad (including the United States, South Africa and Germany) have assessed the more than three and a half million photos taken by camera traps in recent years. With this they have made an essential contribution to increasing knowledge in the field of behavior of the large ungulates in the Park
Many insights thanks to Snapshot
Snapshot Hoge Veluwe has provided a, albeit still limited, insight into the seasonal preference for habitats and foraging areas per animal species. It has shown us how wildlife responds to recreation, for example after the introduction of the foraging ban or after the declaration of a lockdown due to the COVID-19 pandemic. Without the enthusiastic cooperation of these many thousands of volunteers who have completed over three and a half million photo reviews over the past three and a half years, we would not have been able to gain such insight into animal behavior.
Machine learning
Technology is advancing and, thanks to the reviewed photos and the work of computer experts, an algorithm has been developed for Agouti (the computer system), which can already partially take over the photo-judging task. Due to machine learning, a form of artificial intelligence, human insight is no longer central, but the algorithm itself determines a relationship between input and output of data on the basis of a large data set. The computer now recognizes the most common animal species such as red deer, roe deer, wild boars and mouflons through the photos.
The photos taken by the 50-70 camera traps from now on are no longer loaded into Snapshot Hoge Veluwe via Agouti. The translation of all data in this system is being further developed, so that in the future insight will also be obtained through smart intelligence, for example into group size and behaviour.
The studies into animal behavior at De Hoge Veluwe National Park will of course continue, but all this means that the Snapshot Hoge Veluwe website will soon be closed and can no longer be viewed via Zooniverse.
Many thanks to all volunteers
De Hoge Veluwe National Park is very grateful to all volunteers who have been involved in Snapshot Hoge Veluwe in recent years. Without them, the citizen science project Snapshot Hoge Veluwe would not have been such a great success. And sadly for most of them, this success has helped computers to take over their jobs.