[Analysis and modeling of hyperspectral singularity in rice under Cd pollution]

Area of Science:

• Agricultural Science
• Remote Sensing
• Biochemistry

Context:

• Cadmium (Cd) pollution poses a significant threat to rice production and food safety.
• Accurate and rapid detection of Cd stress in rice is crucial for environmental monitoring and agricultural management.
• Hyperspectral imaging combined with signal processing offers a non-destructive approach to assess plant stress.

Purpose:

• To develop and validate a method for fast and accurate detection of Cd stress levels in rice using hyperspectral data.
• To explore the application of wavelet transform and singularity analysis for identifying Cd pollution in rice.
• To establish a regression model for estimating Cd content in rice based on wavelet energy coefficients.

Summary:

• Hyperspectral reflectance data (350-1300 nm) from rice under natural conditions were analyzed using Daubechies 5 wavelet transform.
• Wavelet energy coefficients, particularly at the fifth decomposition level (d5), showed high correlation with Cd content.
• A regression model using third-level wavelet energy coefficients accurately estimated rice Cd content (R²=0.958, RMSE=0.122).

Impact:

• The study demonstrates the potential of wavelet transform and singularity analysis for effective detection of Cd-induced stress in rice.
• Wavelet energy coefficients provide a reliable means to estimate rice Cd content, aiding in pollution monitoring.
• This approach can serve as a reference for detecting other metal-induced stresses in crops.