Instantaneous Material Classification Using a Polarization-Diverse RMCW LIDAR.

Area of Science:

• Photonics and Material Science
• Optical Sensing Technologies

Background:

• Light Detection and Ranging (LIDAR) sensors can acquire polarimetric information.
• Polarization-diverse receivers enhance LIDAR capabilities for target characterization.

Purpose of the Study:

• To demonstrate a silicon photonic receiver for real-time polarimetric LIDAR data acquisition.
• To utilize this data for material classification via a neural network.

Main Methods:

• Development of a silicon photonic receiver architecture for polarization-diverse LIDAR.
• On-shot-by-shot basis polarimetric data capture and analysis.
• Training a neural network with captured polarimetric data for material classification.

Main Results:

• Instantaneous polarization analysis integrated into the LIDAR point cloud.
• Achieved 85.4% classification accuracy for plastic, wood, concrete, and coated aluminum.
• Demonstrated the effectiveness of polarimetric LIDAR for material identification.

Conclusions:

• Silicon photonic receivers enable advanced polarimetric LIDAR capabilities.
• Polarimetric LIDAR coupled with neural networks offers a powerful tool for material classification.
• This technology has potential applications in remote sensing and material identification.