Dense range map reconstruction from a versatile robotic sensor system with an active trinocular vision and a passive...

Area of Science:

• Robotics
• Computer Vision
• Sensor Fusion

Background:

• 3D perception in robotics requires efficient sensor systems for dense range map generation.
• Conventional active and passive vision systems have limitations in reliability and environmental adaptability.
• Developing robust 3D sensing for robots remains a significant research challenge.

Purpose of the Study:

• To develop an efficient visual sensor system for robotic 3D perception.
• To create a cooperative bidirectional sensor fusion method integrating active and passive vision.
• To enable reliable dense range map acquisition using combined sensor data.

Main Methods:

• A visual sensor system combining active trinocular vision and passive stereo vision was developed.
• A cooperative bidirectional sensor fusion algorithm was designed, with two complementary parts.
• The first part uses intensity images to match laser patterns in stereo images; the second uses dynamic programming for pixel-by-pixel matching between laser patterns.

Main Results:

• The proposed sensor fusion method successfully generates reliable dense range maps.
• The system integrates active and passive vision data for enhanced 3D perception.
• Experimental tests on a robotic system demonstrated the effectiveness of the algorithms across various configurations.

Conclusions:

• The cooperative bidirectional sensor fusion approach enhances the reliability of dense range map generation for robotic 3D perception.
• The integrated active trinocular and passive stereo vision system offers a robust solution for real-world robotic applications.
• The developed fusion algorithms provide a significant advancement in robotic sensing capabilities.