Retrieval of Crop Variables from Proximal Multispectral UAV Image Data Using PROSAIL in Maize Canopy

Area of Science:

• Agricultural Science
• Remote Sensing
• Plant Physiology

Background:

• Accurate crop status mapping is vital for farmers and breeders.
• Radiative transfer models (RTMs) offer superior transferability over regression models for crop monitoring.
• RTM inversion provides physically meaningful outputs reflecting canopy processes.

Purpose of the Study:

• To assess the PROSAIL model's capability in retrieving maize crop variables using high-resolution UAV data.
• To compare the effectiveness of mean reflectance versus pixel-based RTM inversion approaches.
• To evaluate the retrieval accuracy of leaf chlorophyll content (LCC), leaf area index (LAI), and canopy chlorophyll content (CCC).

Main Methods:

• Utilized coupled leaf-canopy RTM PROSAIL with multispectral UAV data (0.015 m resolution).
• Applied RTM inversion to mean reflectance and pixel-based orthomosaic data.
• Implemented vegetation index thresholding to remove soil and shaded pixels for LCC retrieval.

Main Results:

• Promising LCC retrieval accuracy achieved with the mean reflectance approach (RMSE: 4.92 µg/cm² for sweet maize, 3.74 µg/cm² for silage maize).
• LAI retrieval was challenging due to mixed pixels but accurate in early growth stages (RMSE: 0.70 m²/m² sweet maize, 0.61 m²/m² silage maize).
• The pixel-based approach significantly improved CCC retrieval (RMSE decreased from 45.6 to 33.1 µg/m²).

Conclusions:

• High-resolution UAV imagery is highly suitable for accurate LCC retrieval in maize.
• Pixel-based RTM inversion enhances the mapping accuracy of canopy chlorophyll content.
• LAI retrieval accuracy is influenced by canopy geometry and illumination conditions, particularly in later growth stages.