Inversion Studies on the Heavy Metal Content of Farmland Soils Based on Spectroscopic Techniques: A Review

Wei Qiu, Ting Tang, Song He, Zezhong Zheng, Jinhong Lv, Guo Jiacheng, Yan Zeng, Ye‐Xing Lao, Weibin WuAgronomy 2025 · DOI

TL;DR

This review synthesizes the current state of spectroscopic techniques for rapidly and non-destructively monitoring heavy metal (HM) content in farmland soils. It highlights the challenges posed by soil heterogeneity and low HM concentrations, emphasizing the need for advanced spectral inversion methods across various platforms (ground, airborne, satellite) to ensure food safety and environmental protection.

What Problem Does This Paper Address?

Heavy metal pollution in farmland soils is a growing crisis, particularly in China, impacting crop production and food safety. Traditional methods for detecting these metals are often inefficient, costly, and time-consuming, hindering real-time monitoring. The low concentration of many heavy metals in soil makes them difficult to detect and quantify accurately. This paper addresses the need for effective, rapid, and large-scale monitoring solutions by reviewing the application of spectral sensing technology for quantitative inversion of heavy metal content in farmland soils.

Methods

This review focuses on the application of spectral sensing technology, specifically Visible-Near-Infrared (VNIR) and Short-Wave Infrared (SWIR) spectroscopy (350–2500 nm), for heavy metal inversion. The authors discuss:

  • Data Acquisition Platforms: Ground-based (proximal sensing), airborne, and satellite-based spectroscopy are examined for their strengths and weaknesses in spatial coverage, resolution, and cost.
  • Spectral Inversion Mechanisms: The paper delves into how soil heterogeneity (particle size, organic matter, pH, etc.) influences spectral responses and how heavy metals interact with soil components, leading to subtle spectral features.
  • Preprocessing Techniques: Various methods for noise reduction, feature enhancement, and spectral correction (e.g., Savitzky-Golay smoothing, continuum removal, standard normal variate) are discussed to improve signal quality.
  • Feature Extraction: Techniques like Principal Component Analysis (PCA), Uninformative Variable Elimination (UVE), and Successive Projection Algorithm (SPA) are reviewed for dimensionality reduction and identifying key spectral bands.
  • Modeling Methods: Both linear (e.g., PLSR, MLR) and nonlinear (e.g., Random Forest, SVM, Deep Learning) models are analyzed for their effectiveness in predicting HM concentrations.
  • Model Evaluation: Key metrics like R², RMSE, and RPD are discussed for assessing model accuracy and generalizability.

Datasets

The paper is a review and does not present new datasets. It synthesizes findings from over 140 literature sources, with a primary focus on studies conducted in China. The authors note that the reliance on Chinese farmland data introduces geographical limitations.

Key Insights and Ideas

  • Indirect Inversion is Key: Due to the low concentration and weak spectral signatures of heavy metals, direct inversion is often challenging. The paper emphasizes that indirect inversion, leveraging the correlation between heavy metals and soil components like organic matter, clay minerals, and metal oxides, is a more viable approach.
  • Soil Heterogeneity is a Major Hurdle: Variations in soil particle size, organic matter content, pH, and moisture significantly impact spectral responses, making it difficult to develop universally applicable inversion models. This necessitates careful consideration of these factors during sampling and model development.
  • VNIR-SWIR is the Sweet Spot: The 350–2500 nm spectral range (VNIR-SWIR) is highlighted as particularly effective for capturing the subtle spectral features related to heavy metals and their interactions with soil components.
  • No One-Size-Fits-All: The review underscores that different preprocessing, feature extraction, and modeling techniques are optimal for different soil types, heavy metals, and data acquisition platforms. This points to the need for tailored approaches rather than a single universal solution.

Key Results

The review synthesizes numerous studies demonstrating the feasibility of spectral techniques for heavy metal inversion. For instance, studies cited show R² values ranging from 0.62 to 0.94 for predicting heavy metal content in agricultural soils. The paper highlights successful applications in predicting concentrations of various metals like Cd, As, Cr, Cu, Pb, and Hg across different soil types and regions, primarily in China.

Limitations and Future Work

The authors identify several key challenges and suggest future research directions:

  • Model Generalizability: A significant challenge is the lack of universally applicable models due to the diverse nature of farmland soils and varying environmental factors. Future work should focus on developing global spectral libraries and standardized calibration/modeling methods.
  • Regional Bias: The heavy reliance on Chinese data limits the global applicability of current findings. More research is needed from other geographical regions.
  • Automation and Cost: The current automation level for spectral analysis is low, and equipment costs remain high. Improving automation and exploring cost-effective solutions are crucial for wider adoption.
  • Data Fusion: Integrating data from different platforms (ground, airborne, satellite) remains a challenge due to differences in spatial resolution, spectral bands, and acquisition times. Developing effective data fusion algorithms is essential for large-scale, high-precision monitoring.
  • Beyond VNIR-SWIR: Exploring spectral bands beyond the VNIR-SWIR range could potentially yield more information about heavy metals.
  • Advanced Machine Learning: Further exploration of machine learning methods, particularly those adept at handling multivariate and nonlinear problems, is recommended to enhance model accuracy and generalization.

Summary updated

Abstract

In recent years, heavy metal pollution in farmland soil has become a crisis due to human activities or natural impacts, with particular emphasis on cases from China, where this issue is prominent, greatly affecting crop production and food safety. In the context of a low heavy metal (HM) content in farmland soil, which is difficult to monitor in real time, effective and rapid monitoring of soil plays a decisive role in subsequent targeted protection measures. To this end, this paper provides a narrative review of the application of spectral sensing technology on the basis of the quantitative inversion of heavy metal content in farmland soil using different platforms (ground, airborne, and spaceborne). The sensing process evaluates the mechanism by which soil produces different weak spectral features from the perspective of the heterogeneity of farmland soil. Different methods used for the quantitative inversion of heavy metals (by studying the correlation between soil heavy metals and organic matter, clay minerals, metal oxides, crop vegetation index, etc.) and their feasibility were clarified. At the same time, relevant research on key technologies used in various processes—such as follow-up pretreatment, spectral feature extraction, and the establishment of inversion models for spectral data of different farmland soil types—was summarized, with a primary focus on cases in China. Finally, the challenges, applications, and research directions related to heavy metal spectral inversion in farmland soil were discussed.
BibTeX
@inproceedings{qiu2025inversion,
  author = {Wei Qiu and Ting Tang and Song He and Zezhong Zheng and Jinhong Lv and Guo Jiacheng and Yan Zeng and Ye‐Xing Lao and Weibin Wu},
  title = {{Inversion Studies on the Heavy Metal Content of Farmland Soils Based on Spectroscopic Techniques: A Review}},
  booktitle = {Agronomy},
  year = {2025},
  doi = {10.3390/agronomy15071678},
}
Figure