ROI-driven optical compression and prism-grating dispersion enabling ultra-high-speed hyperspectral imaging

Area of Science:

• Optics and Photonics
• Spectroscopy
• Materials Science

Background:

• Traditional push-broom hyperspectral cameras face frame rate limitations due to pixel-velocity mismatch.
• This hinders real-time industrial applications requiring high-speed imaging and analysis.

Purpose of the Study:

• To develop a hyperspectral imaging system overcoming traditional frame rate limitations.
• To enable high-speed, precise material identification for industrial sorting applications.

Main Methods:

• Synergistic co-optimization of sensor region-of-interest (ROI) functionality and prism-grating dispersion.
• Utilizing ROI-controlled readout mode for ultra-high-speed spectral image acquisition (3448 fps).
• Custom-designed prism-grating module to suppress distortions (Keystone, Smile) to sub-pixel levels.

Main Results:

• Achieved ultra-high-speed spectral imaging at 3448 fps with suppressed edge jitter.
• Maintained average spectral resolution of 13 nm, enabling identification of 6-mm-scale targets.
• Successfully discriminated between PS, PE, and PMMA using spectral fingerprint analysis under simulated industrial conditions.

Conclusions:

• The developed system offers a hardware-level solution for real-time hyperspectral imaging in industrial scenarios.
• Demonstrated applicability for precise material discrimination in industrial sorting.
• Overcomes previous frame rate limitations, paving the way for advanced industrial automation.