Sugarcane Nitrogen Concentration and Irrigation Level Prediction Based on UAV Multispectral Imagery

Area of Science:

• Agricultural remote sensing
• Crop physiology
• Precision agriculture

Background:

• Sugarcane growth depends on optimal fertilizer, water, and light.
• Unmanned Aerial Vehicle (UAV)-based multispectral imagery offers rapid, field-scale crop vigor assessment.

Purpose of the Study:

• To evaluate the potential of drone multispectral images for predicting sugarcane canopy nitrogen concentration (CNC) and irrigation levels.
• To establish predictive models for CNC and irrigation using spectral data.

Main Methods:

• Collected multispectral images from a sugarcane field under varying irrigation and fertilizer levels.
• Employed Partial Least Square (PLS), Backpropagation Neural Network (BPNN), and Extreme Learning Machine (ELM) for CNC prediction.
• Utilized Support Vector Machine (SVM) and BPNN for irrigation level classification based on spectral features.

Main Results:

• The PLS model achieved high accuracy (Rv = 0.79, RMSEv = 0.11) for CNC prediction using specific vegetation indices and band reflectance.
• The SVM model demonstrated 80.6% accuracy in classifying irrigation levels.
• Selected indices like SRPI, NPCI, and NGBDI showed strong correlation with CNC.

Conclusions:

• High-resolution drone multispectral imagery provides effective data for sugarcane CNC prediction.
• This technology enables accurate recognition of water irrigation levels in sugarcane crops.
• Optimized spectral indices and machine learning models enhance precision agriculture applications.