What Is The ACCURACY OF DRAGONFLY

De Transcription | Bibliothèque patrimoniale numérique Mines ParisTech
Aller à : navigation, rechercher


Dragonfly is Onit’s slicing-edge computer imaginative and prescient indoor localization expertise primarily based on visual SLAM that provides accurate indoor position and iTagPro product orientation to forklifts, automated guided vehicles (AGV), autonomous cellular robots robots (AMR), robots, drones and every other transferring vehicle and asset. Dragonfly allows RTLS options for analytics, productiveness and safety in GPS-denied environments like warehouses, production plants, factories and so forth.. Dragonfly delivers the X-Y-Z coordinates and 3D orientation of any shifting machine with centimeter-level accuracy, by analyzing in real-time the video stream coming from a regular huge-angle digital camera connected to a small computing unit. Dragonfly represents the state-of-the-artwork for indoor localization technologies at locations where GPS/GNSS cannot be used and it is much more aggressive compared to different indoor localization applied sciences primarily based on LiDAR, Ultra Wide Bandwidth, Wi-Fi, Bluetooth RSS. HOW DOES IT WORK? In the course of the system setup part the extensive-angle digital camera sends the video feed of its surroundings to the computing unit.



The computing unit takes care of extracting the features of the setting, in every of the frames, and creating a 3D map of the atmosphere (which is geo-referenced utilizing a DWG file of the area). During its utilization in manufacturing the large-angle digital camera sends the real-time video feed of its surroundings to the computing unit. The computing unit extracts the options of the atmosphere in every of the frames and examine them with these inside the previously created 3D map of the setting. This process allows Dragonfly to calculate at greater than 30 Hz the X-Y-Z place and orientation in the 3D area of the digital camera (and thus of the mobile asset on which it's mounted). Dragonfly is an correct indoor location system primarily based on computer vision. The situation is computed in actual time utilizing just an on board camera and a computing unit on board of the machine to be tracked, smart key finder thanks to our computer imaginative and prescient algorithm. Computer vision, odometry and synthetic intelligence are used to create an accurate system, with the intention to deliver a exact location for multiple purposes.



It is a wonderful answer for iTagPro product the exact indoor monitoring of forklifts, iTagPro portable AGV, AMR, robots and drones (within the 3D house). Dragonfly is way more aggressive than LiDAR, UWB, radio signal based mostly technologies for which an advert-hoc infrastructure have to be designed, iTagPro product setup, calibrated and maintained for each specific venue. No receivers, no RFID tags, no antennas, no nodes, no magnetic stripes. Nothing has to be deployed via the venue. You need just a digital camera and a computing unit onboard your mobile automobiles. No tech expertise required, no troublesome directions, iTagPro official no want for error-prone and time-consuming calibrations of ad-hoc UWB infrastructure. SLAM know-how is rather more strong to environmental modifications as opposed to LiDAR, which struggles considerably to keep up accuracy in environments by which obstacles change over time. Dragonfly cameras are simpler to calibrate and are more sturdy to changes within the environment. Dragonfly distributed structure makes the answer dependable by eliminating mandatory server that led to SPOF (single points of failures).



This also means that Dragonfly can develop following the dimensions and growth of your fleet of shifting autos. Dragonfly can work completely offline on a computing unit on board of forklift, AVG, AMR, drones, robots or on an on-premise server. Dragonfly means that you can optimize your operations, increasing the productivity and effectiveness of the tracked units. Along with this its aggressive value, makes the ROI higher than every other expertise presently in the marketplace. Enhance the operations because of a real-time visibility of the particular utilization and path of your cell automobiles (like forklifts) to avoid underneath/over utilization and maximize the efficiency of the fleet. Know in actual-time the location of each transferring asset to forestall accidents between human-guided mobile vehicles (corresponding to forklifts) inside warehouses and production services enabling thus V2V (vehicle to car) and V2P (autos to pedestrians) applications for collision-avoidance. Speed up the productivity by tracking the location of every moving asset to not directly know the place of each dealing with unit on the bottom, racks and shelves.