Invasive pests present serious challenges to New Zealand’s native flora and fauna. Remote sensing is increasingly being harnessed to detect and monitor pests, but there remain challenges around effective and efficient monitoring. AI computer vision techniques, such as deep learning, can reduce the processing burden by automating the extraction of useful information from large quantities of remote sensing data.
Brent Martin and others at Manaaki Whenua have applied deep learning to various environmental detection and monitoring problems. In this talk, Brent presents recent progress in predator monitoring through the automation of pest species classification in trail camera images. He also showcases efforts to automatically map and monitor wilding conifers from satellite data, with the goal of being able to monitor the spread (and control) of wildings through regular generation of national infestation maps.
Негізгі бет Harnessing deep learning in the war against pests
No video
Пікірлер