Scanning-based compressive hyperspectral imaging via spectral-coded illumination.

Area of Science:

• Optics and Photonics
• Image Processing
• Spectroscopy

Background:

• Hyperspectral imaging captures detailed spectral information across spatial dimensions.
• Traditional hyperspectral imaging methods can be complex and data-intensive.
• Compressive sensing offers potential for efficient data acquisition.

Purpose of the Study:

• To introduce a novel scanning-based compressive hyperspectral imaging (CHI) method.
• To develop an advanced tensor-based reconstruction algorithm for 3D hyperspectral data.
• To demonstrate superior performance in visual quality and quantitative analysis.

Main Methods:

• Utilizing spectral-coded illumination for efficient spectral modulation.
• Employing point-wise scanning for spatial information acquisition, applicable to optical scanning systems like lidar.
• Proposing a tensor-based joint reconstruction algorithm incorporating spectral correlation and spatial self-similarity.

Main Results:

• Successful recovery of 3D hyperspectral data from compressive sampled data.
• Demonstrated superior performance in visual quality compared to existing methods.
• Validated through both simulated and real experimental data, showing enhanced quantitative analysis.

Conclusions:

• The proposed scanning-based CHI method with spectral-coded illumination is effective.
• The tensor-based reconstruction algorithm significantly improves hyperspectral image recovery.
• This novel approach offers a promising direction for efficient and high-quality hyperspectral imaging.