[Retrieval model for subtle variation of contamination stressed maize chlorophyll using hyperspectral data]

Area of Science:

• Agricultural Science
• Remote Sensing
• Environmental Science

Background:

• Chlorophyll content is a key indicator of plant health, photosynthesis, and stress.
• Assessing chlorophyll content under natural heavy metal contamination is crucial for crop monitoring.
• Previous studies often used artificial lab conditions with high pollutant levels.

Purpose of the Study:

• To investigate the relationship between hyperspectral indices and maize foliar chlorophyll content.
• To evaluate spectral indices for estimating chlorophyll under natural heavy metal stress.
• To develop accurate models for chlorophyll estimation using hyperspectral data and artificial neural networks.

Main Methods:

• Collected hyperspectral data, foliar chlorophyll, and heavy metal content from maize in three natural fields.
• Computed ten spectral indices after continuum removal (400-800 nm).
• Applied stepwise multiple linear regression and Backpropagation Artificial Neural Network (ANN-BP) for chlorophyll estimation.

Main Results:

• Several spectral indices showed strong correlations with foliar chlorophyll content.
• Multiple linear regression achieved a determination coefficient (R2) of 0.7027.
• The optimized ANN-BP model with two hidden layers yielded a high accuracy (R2 = 0.9758).

Conclusions:

• Hyperspectral indices are effective for estimating maize chlorophyll content under natural heavy metal stress.
• ANN-BP models offer superior accuracy for chlorophyll estimation compared to traditional regression methods.
• This approach provides a non-destructive and efficient tool for monitoring crop health in contaminated environments.