[Study on SPAD visualization of pumpkin leaves based on hyperspectral imaging technology]

Area of Science:

• Plant Pathology
• Agricultural Remote Sensing
• Spectroscopy

Background:

• Downy mildew significantly impacts pumpkin yield, necessitating early detection methods.
• Chlorophyll content is a key indicator of plant health and disease stress.
• Visible/near-infrared hyperspectral imaging offers a non-destructive approach to assess plant physiological status.

Purpose of the Study:

• To develop a hyperspectral imaging technique for visualizing SPAD (chlorophyll) in pumpkin leaves.
• To establish a correlation between downy mildew infection levels and spectral data.
• To create a diagnostic tool for rapid downy mildew detection in pumpkin plants.

Main Methods:

• Visible/near-infrared (380-1030 nm) hyperspectral imaging was employed.
• Competitive Adaptive Reweighted Sampling (CARS) selected 10 optimal wavelengths.
• Partial Least Squares Regression (PLSR) built a SPAD prediction model.

Main Results:

• The CARS-PLSR model demonstrated high accuracy (Rc=0.918, Rp=0.881) for SPAD prediction.
• SPAD distribution maps were generated using imaging processing technology.
• Downy mildew infection was successfully diagnosed based on the generated SPAD maps.

Conclusions:

• Hyperspectral imaging combined with CARS-PLSR is effective for SPAD visualization and downy mildew diagnosis in pumpkin leaves.
• This technique provides a reliable method for monitoring plant growth and disease epidemics.
• The study offers a valuable reference for early disease detection in agriculture.