Predicting copper contamination in wheat canopy during the full growth period using hyperspectral data
Guodong Wang1, Qixin Wang1, Zhongliang Su2
1College of Chemical engineering, Qingdao University of Science & Technology, Qingdao, 266042, China.
Summary
This study developed hyperspectral imaging models to accurately estimate copper content in wheat, even under field conditions. These models show promise for real-time agricultural monitoring and ensuring food safety.
Area of Science:
- Agricultural Science
- Environmental Science
- Remote Sensing
Background:
- Accurate heavy metal monitoring in crops is crucial for human health and environmental safety.
- Hyperspectral imaging offers a promising, non-destructive method for assessing plant health and composition.
- Existing hyperspectral methods face challenges adapting from lab to field conditions and varying spectral data applicability.
Purpose of the Study:
- To develop and validate field-based hyperspectral models for predicting copper (Cu) content in wheat canopies.
- To identify sensitive spectral bands and indices correlated with copper accumulation in wheat.
- To assess the applicability of these models across different growth stages under actual agricultural conditions.
Main Methods:
- Field experiments were conducted throughout the entire growth period of wheat.
- Wheat canopy hyperspectral data were collected and processed using the first derivative method.
- Sensitive spectral bands and indices were screened based on their correlation with measured copper content.
- Predictive models (linear regression, NDVI/SIPI, Rg, W728, W741, W480, multiple bands) were built and validated for different growth stages.
Main Results:
- Copper content in wheat correlated positively with soil copper levels.
- Subtle differences in wheat canopy spectral reflectance were observed under varying soil copper concentrations.
- Optimal predictive models varied by growth stage, with R-squared values ranging from 0.548 to 0.868.
- Models based on specific indices (NDVI/SIPI) and wavelengths (W728, W741, W480) demonstrated strong predictive power.
Conclusions:
- Hyperspectral modeling shows significant potential for estimating wheat copper content in field settings.
- The developed models are suitable for application during different wheat growth stages.
- This approach offers a viable tool for real-time monitoring in agricultural production, contributing to food safety and environmental protection.
