Towards Low-Cost Hyperspectral Single-Pixel Imaging for Plant Phenotyping

Area of Science:

• Optics and Photonics
• Remote Sensing
• Plant Science

Background:

• Hyperspectral imaging (HSI) offers valuable data for various applications but remains costly for widespread use.
• Existing HSI technologies are often expensive, limiting their adoption in moderate to low-cost indoor and outdoor settings.
• There is a need for affordable HSI solutions for applications like plant phenotyping and education.

Purpose of the Study:

• To develop and validate a low-cost hyperspectral imaging prototype.
• To enable metrologically validated reflectance measurements with minimal computational load.
• To demonstrate the capability of discriminating plant species and identifying areas within leaves.

Main Methods:

• Utilized compressive sensing methods with a single-pixel imaging setup.
• Generated projected patterns on a Fourier basis to reduce acquisition and calculation times.
• Developed and tested a low-cost, moderate-flow prototype in a laboratory setting.

Main Results:

• Achieved metrologically validated reflectance measurements with a minimal computational workload.
• Successfully discriminated plant species from surrounding scene elements.
• Identified biologically contrasted areas within plant leaves.

Conclusions:

• The developed prototype provides a low-cost solution for hyperspectral data acquisition.
• This technology democratizes access to plant phenotyping and educational tools.
• The system demonstrates the potential for affordable, high-performance spectral analysis.