At-Sensor Radiometric Correction of a Multispectral Camera (RedEdge) for sUAS Vegetation Mapping

Area of Science:

• Remote Sensing
• Agricultural Science
• Photogrammetry

Background:

• Small unmanned aircraft systems (sUAS) with advanced sensors are increasingly used for quantitative applications.
• MicaSense RedEdge cameras are popular in agriculture for crop health assessment, relying on calibrated reflectance panels (CRP) for surface reflectance extraction.

Purpose of the Study:

• To evaluate the performance of a Matrace100/RedEdge-M camera system in extracting surface reflectance orthoimages.
• To develop an improved radiometric correction model for enhanced accuracy and stability.
• To propose a new vegetation index for more reliable crop health assessment.

Main Methods:

• Developed an at-sensor radiometric correction model integrating CRP and a Downwelling Light Sensor (DLS).
• Conducted multiple flights and field experiments at three vegetated sites.
• Proposed and validated a red edge Normalized Difference Vegetation Index (ReNDVI).

Main Results:

• The standard CRP-only correction showed limitations, particularly in the Near-Infrared (NIR) band and under variable weather conditions.
• The proposed model significantly reduced local impacts on extracted surface reflectance.
• Normalized Difference Vegetation Index (NDVI) exhibited overestimation and saturation in vegetated areas; ReNDVI offered improved performance.

Conclusions:

• The integrated radiometric correction model enhances the reliability of drone-based surface reflectance data.
• ReNDVI is a more robust index for assessing vegetation health across diverse conditions compared to traditional NDVI.
• This research contributes to more accurate and stable drone applications in agroindustry.