00:00The number of animals living in Germany's national parks is large.
00:04But how are their populations changing and how is climate change affecting them?
00:09To answer these questions, researchers would need to comb through massive amounts of data,
00:14millions of images and thousands of hours of audio files.
00:17It's too much for humans to do, but light work for AI.
00:25We would need at least 20 of me just classifying images.
00:32Isabel Jorczyk and her colleague Marco Heurich are out in the Bavarian Forest National Park,
00:37checking camera traps, microphones and sensors recording data around the clock.
00:42Let's see what's on here.
00:46A bit of brute force.
00:48We have the camera traps and they have a sensor which triggers when something warm and quick is passing by,
00:56so basically ground-dwelling animals.
00:58We have red deer and road deer mostly here and we mostly see red deer and also wild boars.
01:04What is rare are lynx and wolves.
01:07But we also see a lot of birds that are landing in front of the camera and can trigger the
01:11camera.
01:13Back at the office, it's over to the AI, which processes the data with various algorithms.
01:19First, any photos containing humans are blurred for privacy.
01:23Then, one program identifies the species while others determine the age and biological sex.
01:30Then it's back to the humans, who double-check for any mistakes.
01:33After all, AI is far from perfect.
01:37So even though sometimes the AI is really confident in its guess, a human will know that this is not
01:43true,
01:44because of course the AI can only classify what it was trained on.
01:48The team is also testing new sound analysis methods.
01:53Acoustic monitoring uses microphones to record animal noises, including ultrasound.
02:03This allows different species to be identified by their calls, without human presence distorting the results.
02:11These AI tools being tested all over Germany could eventually be used across the world,
02:16to better protect national parks and the animals within them.
Comments