Zhu et al. (2025) Applications of Polarization Spectroscopy in Agricultural Engineering: A Comprehensive Review

Identification

• Journal: Agriculture
• Year: 2025
• Date: 2025-12-09
• Authors: Wenjing Zhu, Liangliang Zhai, Wenhao Du, Li Xiao, Zhiqiao Gao, Huan Wang, Yang Li
• DOI: 10.3390/agriculture15242546

Research Groups

• School of Agricultural Engineering, Jiangsu University, Zhenjiang, China
• High-Tech Key Laboratory of Agricultural Equipment and Intelligence, Zhenjiang, China
• Weichai Lovol Intelligent Agricultural Technology Co., Ltd., Weifang, China

Short Summary

This comprehensive review synthesizes the principles and applications of polarization spectroscopy analysis (PSA) in agricultural engineering, demonstrating its advantages over conventional methods for non-destructive testing across various agricultural materials and environments. It highlights PSA's potential to enhance precision agriculture through improved monitoring of crop health, pest/disease detection, quality assessment, and environmental evaluation.

Objective

• To provide a comprehensive overview of the principles, advantages, and applications of polarization spectroscopy analysis (PSA) in agricultural engineering, specifically focusing on non-destructive testing tasks across multiple scales and agricultural scenarios.

Study Configuration

• Spatial Scale: Laboratory-level (leaf, seed, fruit microstructure), field-level (crop canopy, soil), and remote sensing (UAV, satellite platforms, large-scale farmland, aquaculture environments).
• Temporal Scale: Real-time monitoring, early detection of stress factors, various crop growth stages, and seasonal dynamics.

Methodology and Data

• Models used:
  • PROSPECT model
  • PROPOLAR model
  • Bidirectional Polarization Distribution Function (BPDF) models (Nadal–Bréon, Litvinov, Xie–Cheng)
  • Partial Least Squares Discriminant Analysis (PLS-DA)
  • Support Vector Regression (SVR)
  • Gaussian Process Regression (GPR)
  • Partial Least Squares Regression (PLSR)
  • Random Forest (RF)
  • Support Vector Machine (SVM)
  • K-Nearest Neighbors (KNN)
  • Back-Propagation Neural Network (BPNN)
  • Radial Basis Function Neural Network (RBFNN)
  • YOLOv5 (improved)
  • ResNet-G18 (ResNet-18 + Ghost)
  • Convolutional Neural Networks (CNN)
  • Deep Neural Networks (DNNs)
  • Water Cloud Model (WCM)
  • Integral Equation Model (IEM)
  • Stokes parameters, Degree of Polarization (DoP), Angle of Polarization (AoP), Degree of Linear Polarization (DoLP)
• Data sources:
  • Polarization spectroscopy and imaging (including multi-angle and multi-band)
  • Hyperspectral imaging and spectroscopy
  • Multispectral imaging and spectroscopy
  • RGB imaging
  • Fluorescence imaging and spectroscopy
  • LiDAR (laser polarization imaging, polarized active-imaging LiDAR, polarization-sensitive LiDAR, shipborne polarized marine LiDAR)
  • Synthetic Aperture Radar (SAR) data (C-band Sentinel-1, X-band COSMO SkyMed, TerraSAR-X dual-polarization)
  • Satellite remote sensing (e.g., GF-1)
  • UAV-mounted sensors
  • Ground-based observations
  • Laboratory measurements
  • Meteorological data

Main Results

• Polarization spectroscopy analysis (PSA) significantly enhances non-destructive testing in agriculture by providing structural information and robustness in complex environments, outperforming conventional optical methods.
• Crop Health and Disease Detection: PSA improves chlorophyll content estimation (e.g., R² up to 0.948 for water bodies, RMSE reduced by approximately 25% for tea trees), enables early detection of water stress (R = 0.85, mean squared error = 0.45%), and enhances nitrogen estimation (e.g., R² = 0.82 for leaf nitrogen concentration, 1.19% reduction in canopy nitrogen content RMSECV). It also facilitates pre-symptomatic disease identification (e.g., HLB in citrus) and classification of fungal spores (86.67% accuracy).
• Agricultural Product and Seed Quality: PSA is effective for monitoring egg freshness (92% accuracy), differentiating crops from weeds (98.2% accuracy), detecting bruises in nectarines (96.21% accuracy), and assessing seed vitality (over 90% accuracy, up to 98.25% with hyperspectral fusion).
• Soil and Environmental Monitoring: PSA aids in accurate soil moisture estimation (e.g., RMSE of 0.05 cubic meters per cubic meter, R = 0.69; R² up to 0.98 for polarization degree vs. humidity), enhances heavy metal detection by reducing background interference, and improves saline-alkaline soil monitoring (prediction accuracy improved by 3.06% to 19.75%).
• Underwater Research: Polarization differential imaging and recovery methods improve image clarity, suppress scattering, and extend detection range in complex underwater agricultural environments, supporting biological monitoring and behavioral analysis (e.g., grass carp feeding).
• Technology Fusion: The integration of polarization with hyperspectral, multispectral, and fluorescence technologies, combined with machine learning and deep learning, consistently improves detection accuracy, robustness, and information richness across diverse agricultural applications.

Contributions

• Integrates laboratory, field, and remote-sensing polarimetric spectral analysis into a unified cross-scale framework, linking fundamental optical mechanisms with practical agricultural monitoring scenarios.
• Places special emphasis on non-destructive quality monitoring, including nutrient assessment, stress detection, disease identification, seed viability evaluation, fruit and product quality inspection, and soil property estimation.
• Summarizes advances in multi-modal fusion, combining polarization with hyperspectral, multispectral, RGB, and fluorescence imaging, and highlights the role of machine learning and deep learning in enhancing detection accuracy.
• Provides an engineering-oriented perspective by reviewing instrumentation, measurement systems, and deployment platforms ranging from laboratory spectrometers to field polarimeters, UAV payloads, and satellite sensors.

Funding

• Key Laboratory of Modern Agricultural Equipment and Technology Project (MAET202322)
• National Natural Science Foundation of China (32071905)
• National Natural Science Foundation of China (61901194)

Citation

@article{Zhu2025Applications,
  author = {Zhu, Wenjing and Zhai, Liangliang and Du, Wenhao and Xiao, Li and Gao, Zhiqiao and Wang, Huan and Li, Yang},
  title = {Applications of Polarization Spectroscopy in Agricultural Engineering: A Comprehensive Review},
  journal = {Agriculture},
  year = {2025},
  doi = {10.3390/agriculture15242546},
  url = {https://doi.org/10.3390/agriculture15242546}
}