Khorram et al. (2025) Harnessing hyperspectral imaging and machine learning to enhance salinity stress detection in canola

Identification

• Journal: Computers and Electronics in Agriculture
• Year: 2025
• Date: 2025-12-08
• Authors: Mahdis Khorram, Saurav Kumar, Rajan Shrestha, Qingwu Xue, Andrea Leiva Soto, Santosh S. Palmate, Girisha Ganjegunte
• DOI: 10.1016/j.compag.2025.111280

Research Groups

• School of Sustainable Engineering and the Built Environment, Arizona State University, USA
• Texas A&M AgriLife Research and Extension Center at Amarillo, USA
• Texas A&M AgriLife Research and Extension Center at El Paso, USA
• Department of Biological and Agricultural Engineering, Texas A&M University, USA

Short Summary

This study applied hyperspectral imaging and machine learning to detect and classify salinity stress in canola, achieving 82.61% accuracy with a ridge classifier using a reduced set of 15 spectral features and novel vegetation indices.

Objective

• To investigate the application of hyperspectral imaging and machine learning techniques for detecting and classifying salinity stress in canola (Brassica napus L.).

Study Configuration

• Spatial Scale: Individual canola plants (Brassica napus L.).
• Temporal Scale: Not explicitly stated, but implies a period sufficient for salinity stress development and detection.

Methodology and Data

• Models used: Ridge classifier model.
• Data sources: Hyperspectral imaging, spectral bands, vegetation indices.

Main Results

• Spectral signature analysis revealed significant changes in reflectance patterns for wavelengths exceeding 740 nm, corresponding to the near-infrared (NIR) region.
• Two novel vegetation indices were developed specifically for salinity stress detection in canola.
• The final ridge classifier model achieved 82.61 % classification accuracy on the test set using only 15 features, a substantial reduction from an initial 331 features.
• The most effective features primarily spanned wavelengths corresponding to Sentinel-2A bands, with notable exceptions at 405.04 nm and 983.96 nm.
• Sentinel-2 bands B1, B5, B6, B7, and B9 may have limited efficacy in identifying salinity stress in canola.

Contributions

• Developed two novel vegetation indices tailored for salinity stress detection in canola.
• Demonstrated a highly accurate (82.61 %) and efficient (15 features) machine learning model for classifying six levels of salinity stress in canola using hyperspectral data.
• Identified specific effective spectral regions and wavelengths for canola salinity stress detection, including those outside Sentinel-2's capabilities, highlighting the potential for crop-specific optimization.
• Contributes to the growing body of knowledge on non-invasive crop stress detection and precision agriculture practices for sustainable agriculture.

Funding

• Not explicitly stated in the provided text.

Citation

@article{Khorram2025Harnessing,
  author = {Khorram, Mahdis and Kumar, Saurav and Shrestha, Rajan and Xue, Qingwu and Soto, Andrea Leiva and Palmate, Santosh S. and Ganjegunte, Girisha},
  title = {Harnessing hyperspectral imaging and machine learning to enhance salinity stress detection in canola},
  journal = {Computers and Electronics in Agriculture},
  year = {2025},
  doi = {10.1016/j.compag.2025.111280},
  url = {https://doi.org/10.1016/j.compag.2025.111280}
}