Research on a UAV spray system combined with grid atomized droplets.

Area of Science:

• Agricultural Engineering
• Precision Agriculture
• Environmental Science

Background:

• Unmanned Aerial Vehicles (UAVs) offer potential for crop protection in China's mountainous orchards.
• Challenges include pesticide droplet drift and limited solution capacity, hindering efficiency and increasing usage.
• Optimizing UAV spraying technology is crucial for enhanced crop protection and reduced environmental impact.

Purpose of the Study:

• To propose and evaluate a novel plant protection UAV spraying method.
• To investigate the effects of operational parameters on droplet deposition and drift.
• To develop a machine learning model for predicting UAV spraying performance.

Main Methods:

• Conducted UAV spray and grid impact tests on three citrus tree types (traditional, dwarf, hedged).
• Analyzed droplet deposition and drift rates under varying particle sizes and UAV altitudes.
• Utilized machine learning (BP, REGRESS, ELM, RBFNN) to model and predict spraying effects.

Main Results:

• Dwarf and hedged trees showed significantly higher droplet deposition than traditional trees.
• Fixed-height grids improved deposition on dense and trellised trees but reduced it on dwarf trees.
• Sampling point height most influenced deposition; wind direction distance most influenced drift.
• ELM model achieved the highest prediction accuracy for droplet drift rate (R² > 0.95, lowest RMSE).

Conclusions:

• The proposed UAV spraying method, combined with optimized operational parameters, can enhance deposition and reduce drift.
• Machine learning, particularly ELM, provides accurate prediction of spraying effects, aiding technology optimization.
• Tailoring spraying strategies to different tree architectures is essential for maximizing efficiency in orchards.