Biomimetic Active Stereo Camera System with Variable FOV

Area of Science:

• Robotics
• Computer Vision
• Biomimetics

Background:

• Active stereo vision systems face challenges with dynamic fields of view (FOVs).
• Existing calibration methods struggle to adapt to real-time FOV changes.

Purpose of the Study:

• To develop a novel dynamic calibration method for active stereo vision systems.
• To enable accurate depth information acquisition during real-time FOV adjustments.
• To create a bionic active camera system mimicking biological eye movements.

Main Methods:

• Integrating static calibration with dynamic updates of extrinsic parameters.
• Leveraging relative pose adjustments between rotation axis and cameras for real-time updates.
• Developing a two-Degrees-of-Freedom (DOF) bionic active camera system.

Main Results:

• The dynamic calibration method effectively updates extrinsic parameters in real-time.
• Epipolar rectification is facilitated during FOV changes.
• High accuracy in extrinsic parameter calculations was achieved during FOV adjustments.

Conclusions:

• The proposed dynamic calibration method is effective for active stereo vision systems with varying FOVs.
• The bionic active camera system demonstrates the practical application of the calibration method.
• This approach enhances the precision of depth information acquisition in dynamic visual environments.