Hardware-based LiDAR and imaging fusion sensing system

Area of Science:

• Robotics and Autonomous Systems
• Sensor Fusion
• Computer Vision

Background:

• Intelligent unmanned systems rely on fused sensor data for enhanced performance.
• Current data-level fusion methods (early, deep, late) face spatial and temporal alignment challenges due to independent sensor operation.
• Effective fusion of LiDAR and multispectral imaging is crucial for advanced perception.

Purpose of the Study:

• To propose and validate a novel hardware-based fusion sensing system integrating LiDAR and multispectral imaging.
• To overcome the limitations of existing data-level fusion techniques.
• To enable high-quality, spatially and temporally synchronized fusion of LiDAR point clouds and multispectral imagery.

Main Methods:

• Development of a hardware-based fusion sensing system integrating LiDAR and multispectral imaging sensors.
• Implementation of a common-path optical configuration for inherent spatial alignment.
• Integration of dedicated hardware circuitry for precise temporal synchronization.
• Conducting outdoor field experiments to evaluate fusion performance.

Main Results:

• The proposed system successfully achieved high-quality fusion of LiDAR point clouds and multispectral imagery.
• Demonstrated effective spatial alignment through the common-path optical design.
• Verified precise temporal synchronization via dedicated hardware circuitry.
• Confirmed the hardware-based fusion approach's effectiveness in real-world conditions.

Conclusions:

• The hardware-based fusion sensing system effectively integrates LiDAR and multispectral imaging data.
• This approach overcomes critical spatial and temporal alignment issues inherent in data-level fusion.
• The system shows significant potential for advancing intelligent unmanned systems through improved sensor fusion.