A cooperative scheme for late leaf spot estimation in peanut using UAV multispectral images

Area of Science:

• Agricultural Science
• Remote Sensing
• Plant Pathology

Background:

• Late leaf spot (LLS) is a major threat to peanut production in Australia.
• Unmanned aerial vehicles (UAVs) offer potential for crop disease monitoring.
• Existing UAV methods using mean or threshold values may not fully represent plot-level disease distribution.

Purpose of the Study:

• To develop and evaluate novel UAV-based methods for estimating LLS disease severity in peanuts.
• To compare the performance of measurement index (MI) and coefficient of variation (CV) methods against traditional approaches.
• To propose an integrated cooperative scheme for automatic LLS disease estimation.

Main Methods:

• Investigated relationships between UAV-based multispectral vegetation indices (VIs) and LLS disease scores.
• Developed and applied MI and CV methods for LLS estimation.
• Compared MI, CV, mean, and threshold methods using peanut field data.
• Proposed a cooperative scheme combining MI, CV, and mean methods.

Main Results:

• The MI-based method showed the highest coefficient of determination and lowest error for most VIs.
• The CV-based method performed best for the simple ratio (SR) index.
• Both MI and CV methods demonstrated improved disease distribution capture compared to mean/threshold methods.
• The cooperative scheme proved effective for automatic LLS estimation.

Conclusions:

• MI and CV methods offer enhanced accuracy and detail for UAV-based LLS disease estimation in peanuts.
• A cooperative scheme integrating MI, CV, and mean methods provides a robust solution for automated disease monitoring.
• These findings can improve precision agriculture practices for peanut cultivation.