Get the latest tech news
WALDO: Whereabouts Ascertainment for Low-Lying Detectable Objects
Whereabouts Ascertainment for Low-lying Detectable Objects. The SOTA in FOSS AI for drones! - stephansturges/WALDO
The basic model shared here, which is the only one published as FOSS at the moment, is capable of detecting these classes of items in overhead images ranging in altitude from about 30 feet to satellite imagery with a resolution of 50cm per pixel or better. 🧍'gastank'--> cylindrical tanks such as butane tanks and gas expansion tanks, or grain silos 🫙'digger' --> all kinds of construction vehicles, including tractors and construction gear 🚜'container' --> shipping containers, including on the back of an articulated truck'bus' --> a bus 🚌'u_pole' --> utility poles, power poles, anything thin and sticking up that you should avoid with a plane 🎏'boat' --> boats 🚢'bike' --> bikes, mopeds, motorbikes, all things with 2 wheels 🚲'smoke' --> smoke and fire 🔥🔥🔥'solarpanels' --> solar panels'arm/mil' --> this class detects certain types of armored vehicles (very unreliable for now, don't use it yet)'plane' --> planes (very unreliable for now, probably not worth using yet) Some of the network that is in this repo is very large, and is meant to be run on an inference server, and some are made for embedding on tiny edge devices... take a look around and find one that works for you!
Or read this on Hacker News