Film, Identify, Track, Tag, Sense, Fly

Опубликовано 6 марта 2014, 23:15

This research from the Microsoft Research Computational Ecology and Environmental Sciences group aims to dramatically expand the amount and kind of data we can gather from the natural world. Zootracer uses vision and machine learning to track arbitrary objects from video. Designed to assist environmental scientists, Zootracer, a tool for general use, is complemented by Mataki, an unprecedentedly cheap, light (seven grams), and reprogrammable GPS tracking and sensing device. Uniquely, Mataki has peer-to-peer data sharing and, hence, data retrieval that can be achieved on entire collections of device-monitored animals. The research also uses an unmanned aerial drone with an onboard camera to follow coordinates broadcast by a Mataki device attached to an animal.

research.microsoft.com