Evaluating Sensor Fusion and Flight Parameters for Enhanced Plant Height Measurement in Dry Peas

Area of Science:

• Agricultural Science
• Remote Sensing
• Plant Breeding

Background:

• Plant height is crucial for assessing lodging, drought, and stress tolerance in crops.
• Traditional plant height measurements are labor-intensive, costly, and prone to errors.
• Existing Unmanned Aerial System (UAS) technologies are primarily tested on erect plants, necessitating research on prostrate crops like dry peas.

Purpose of the Study:

• To compare LiDAR, RGB, and multispectral sensors for accurate dry pea plant height estimation.
• To identify optimal flight configurations (altitude, speed, overlap) for UAS-based measurements.
• To evaluate the impact of sensor fusion on plant height accuracy.

Main Methods:

• Conducted comparative analysis of LiDAR, RGB, and multispectral sensors on dry pea plots.
• Varied flight parameters including altitude, speed, and image overlap.
• Utilized sensor fusion by integrating LiDAR's Digital Terrain Model (DTM) with RGB and multispectral Digital Surface Models (DSMs).

Main Results:

• Higher flight altitudes and increased image overlap generally improved accuracy across all sensors.
• Despite underestimation at higher altitudes, error metrics (RMSE, MAE) showed no significant difference, suggesting cost-efficiency.
• Sensor fusion did not yield significantly better results than individual sensors, though LiDAR offered the highest accuracy in some cases.

Conclusions:

• UAS-based sensors, particularly LiDAR, can accurately estimate dry pea plant height.
• Optimized flight parameters and sensor choice are key for efficient and accurate data collection.
• Future research should integrate machine learning with LiDAR for enhanced height estimation across diverse canopy structures.