We develop 3D Location Monitoring (digital twin) and visual AI ourselves and apply them to video security and automatic control
We build a 3D virtual space that replicates the real one, and move the people and vehicles captured by the cameras into that space in real time
In normal operation XRTS shows only a background image with no people in it in the 3D space, and only the site where an abnormal event occurred is shown as live video, with the camera viewpoint and the viewer's viewpoint aligned
XRTS-based digital twin video security solution (AI Korea 2022, BEXCO)
XRTS-based video security solution for privacy protection (AI Expo 2024, COEX)
We detect moving objects such as people and vehicles in images and video, and estimate human posture
Abnormal events are detected and alerted in real time. Which features apply depends on the product and the site configuration
Virtual FenceCounts the objects that cross a line
Intrusion DetectionDetects an object entering a surveillance area
Loitering DetectionDetects an object staying and moving within a restricted area
Fall-from-Height DetectionDetects a person falling from height in a surveillance area
Collapse DetectionDetects a person who has collapsed
Caught-in DetectionDetects a body being caught in machinery
Privacy ProtectionDisplays people as skeletons or avatars to hide personally identifiable information (optional)
Facial RecognitionControls entry using facial information
Person TrackingTracks the location of a specified person in real time
Location TrackingConverts geospatial information into 3D to show object locationsWe track the movement paths of people and vehicles across multiple cameras in real time and show them on a flat map or a GIS-based 3D view (registered patents 10-2152318 · 10-2152319)
People in the video are displayed in different ways to match the security requirements of the site. De-identified display is an optional feature

Building and facility structures and devices such as HVAC units are reproduced in 3D space for monitoring and remote control


