Automated BRDF Measurement for Aerospace Materials and 1D-CNN-Based Estimation of Mixed-Material Composition

Area of Science:

• Space Surveillance and Object Identification
• Optical Physics and Material Science
• Artificial Intelligence in Aerospace

Background:

• Space-based optical surveillance is crucial but limited by diffraction for distant objects.
• Traditional image-based recognition fails due to low spatial resolution.
• Bidirectional Reflectance Distribution Function (BRDF) offers superior material identification through 4D features.

Purpose of the Study:

• To develop an automated system for measuring BRDF of aerospace materials.
• To investigate BRDF properties of mixed-material surfaces.
• To create a predictive model for material composition ratios using BRDF data.

Main Methods:

• Automated BRDF measurement system development.
• Characterization of typical aerospace materials and mixed surfaces.
• Construction of a 1D Convolutional Neural Network (1D-CNN) for material ratio prediction.

Main Results:

• The 1D-CNN model effectively extracts BRDF curve features.
• Maximum relative percentage error of 6.21% achieved in composition ratio prediction.
• Prediction accuracy for mixed-material composition ratios consistently exceeded 93.79%.

Conclusions:

• BRDF signatures offer an effective approach for Space Object Identification.
• The proposed 1D-CNN model provides efficient, robust, and less complex material identification.
• This research supports space-based material identification and intelligent space situational awareness.