Spatio-Temporal Calibration of Multiple Kinect Cameras Using 3D Human Pose

Area of Science:

• Computer Vision
• Robotics
• Human-Computer Interaction

Background:

• RGB and depth cameras capture 3D human pose and motion using skeletal structures.
• Single-camera tracking suffers from noise and occlusions.
• Multi-camera systems enhance coverage but require spatio-temporal calibration.

Purpose of the Study:

• To develop an on-the-fly spatio-temporal calibration method for multi-camera systems.
• To eliminate the need for specialized calibration devices and equipment.
• To enable accurate 3D human pose and motion tracking using only human movement.

Main Methods:

• Introduced a novel approach for on-the-fly spatio-temporal calibration.
• Utilized human body motion within the scene for calibration.
• Implemented and evaluated the system using Microsoft Azure Kinect.

Main Results:

• The proposed method achieves accuracy and robustness comparable to state-of-the-art calibration techniques.
• Demonstrated effective spatio-temporal calibration without external devices.
• Successfully enabled 3D human tracking in multi-camera setups.

Conclusions:

• On-the-fly spatio-temporal calibration using human motion is feasible and effective.
• This approach simplifies multi-camera 3D human tracking system setup.
• The method offers a robust and accurate alternative to traditional calibration procedures.