AI-enabled real-time dual-comb molecular fingerprint imaging

Area of Science:

• Spectroscopy
• Optical Imaging
• Infrared Technology

Background:

• Hyperspectral imaging offers spatially resolved spectral data.
• Dual-frequency combs enable ultra-high spectral resolution in scan-free imaging via heterodyne detection.

Purpose of the Study:

• To demonstrate dual-comb hyperspectral imaging with an uncooled near-to-mid-infrared detector.
• To achieve high-speed, real-time imaging of molecular signatures.

Main Methods:

• Utilized dual-frequency combs for active illumination.
• Employed a detector array with a high frame rate for heterodyne detection.
• Applied artificial intelligence for data reduction and gas concentration analysis.

Main Results:

• Achieved 10 Hz acquisition rate with 30 spectral channels across 16,384 pixels.
• Demonstrated real-time imaging of gas concentration using molecular absorption signatures.
• Leveraged detector sensitivity from 1 to 5 µm wavelength.

Conclusions:

• This work establishes a foundation for real-time, versatile infrared hyperspectral imaging.
• The system enables high-temporal-resolution imaging of molecular fingerprint signatures.
• The integration of AI facilitates efficient data processing and analysis.