Communications of the ACM
Drones Survey African Wildlife
Scientists are using drone flights and automated image analysis to develop a new approach to counting animals in the wild.
Credit: gateway-africa.com
Scientists funded by the Swiss National Science Foundation (SNSF) are using drone flights and automated image analysis to develop a new approach to counting animals in the wild.
The new technique enables fast and accurate counting of gnu, oryx, and other large mammals living in wildlife reserves.
The drones remotely photograph wilderness areas, then the images are analyzed using object-recognition software and verified by humans. The drones enable researchers to study vast areas, with more than 150 images captured for each square kilometer.
The researchers use deep learning to analyze this mass of raw visual data, eliminating most images containing no wildlife; in other images, the algorithm highlights the patterns most likely to be animals.
The team trained the artificial intelligence system using an international crowdsourcing campaign in which volunteers tracked animals in thousands of aerial photos of the savanna taken from a Namibian wildlife reserve.
From Swiss National Science Foundation
View Full Article
Abstracts Copyright © 2018 Information Inc., Bethesda, Maryland, USA
No entries found