2023-07-15 12:30:58
Already behind a system to count elephants automatically, the non-profit organization WildEye has developed an open-source application to process data from camera traps.
You will also be interested
[EN VIDÉO] Wild animals in the face of global warming Even very slight changes in the climate can strongly affect the biodiversity of an area, even…
68% of vertebrate populations have disappeared in less than 50 years according to and WWF… To follow the evolution of fauna or a process of reintroduction, scientists specialized in conservation biology mainly rely on data from camera traps. The problem is that, in addition to numerous shots triggered by the wind or other disturbances, there are thousands of images to process, annotate, classify, with potential sources of error and a considerable waste of time. This is where artificial intelligence can be of great help, as with Trap Taggera solution developed by the non-profit organization WildEye.
I find out how this solution works!
Three complementary artificial intelligence algorithms
Developed with the University of Oxford Wildlife Conservation Research Unit (WildCRU), TrapTagger is an open source web application for processing camera trap data. It uses three artificial intelligence algorithms for this: one detects and counts the animals in the images, another classifies the animal species, while the third calculates the similarity between the animal coat patterns, which facilitates identification. individuals. In addition, annotation interfaces allow scientists to choose the observation approach that suits them, automatic or manual, depending on the desired level of detail.
A 99% accuracy rate
The result is impressive as TrapTagger detects, counts, classifies and individually identifies animals at 99% accuracy rates, reducing data processing time and costs by 98%. The application is also open source and totally free. It already processes more than two million images per month for some forty organizations which can thus, if they wish, share their information and be more precise in the statistical data. Of course, TrapTagger guarantees their complete confidentiality so as not to reveal the location of camera traps or to prevent images of vulnerable species from falling into the wrong hands.
To find out, click here!
Article written in partnership with Hello Future, Orange’s research and innovation site
1689431894
#Improving #wildlife #conservation #artificial #intelligence