This technology is a system and method that maintains vision tracking accuracy for a target by detecting the real-time movement of a mobile robot's first body via indoor GPS, using this as a feedforward signal, and inputting it along with a feedback signal from a vision sensor into a controller to actively drive a second body, which serves as a camera mount.
When the driving unit of a mobile robot rotates or vibrates, the vision sensor moves along with the robot body, causing the target to move out of the recognition range or resulting in motion blur in the video signal, which leads to reduced recognition rates and accuracy.
This technology proposes an active vision tracking system that acquires movement and rotation information of the first body through indoor GPS triangulation and compensates for it by driving the second body in the opposite direction of the first body's movement via a controller. By using inertial sensors and encoders in combination to correct motion detection errors, it can perform precise tracking. It can be applied to indoor surveillance robots, logistics automation equipment, and precision imaging devices, ensuring stable video tracking even while in motion through precise, indoor GPS-based compensation control.
N/A
