An Automated Workflow for Pixel-Level BRDF Extraction Using UAV-Based Multispectral Images

Bidirectional Reflectance Distribution Function (BRDF) plays a vital role in quantitative remote sensing. Recently, UAV has gradually emerged as the leading choice for BRDF acquirement. However, challenges remain in extracting usable BRDF data from UAV multispectral images (MSIs), including labor-intensive and limited accuracy. To tackle these challenges, an automated workflow for pixel-level BRDF extraction is proposed in this paper, comprising two main stages: 3D reconstruction and back projection. Pixel-level accuracy is achieved through 3D reconstruction, enhanced by the integration of commercial software for streamlined automation. Back projection is critical for precise location. Experiments were carried out for both selected ROIs and all pixel areas, validating the efficacy of our proposed approach.