Multispectral imaging through scattering media and around corners via spectral component separation

Area of Science:

• Optics and Photonics
• Image Processing
• Spectroscopy

Background:

• Current imaging techniques struggle with multispectral imaging in scattering media, yielding poor spectral recovery.
• Existing methods often produce grayscale images and fail to resolve complex multispectral speckle patterns.

Purpose of the Study:

• To develop a robust method for high-quality multispectral imaging through scattering media and around corners.
• To establish a spectral component separation model that addresses limitations of existing approaches.

Main Methods:

• Constructing a multispectral speckle simplex using uncorrelated speckles from different wavelengths.
• Employing spectral intensity modulation and a joint-solving mechanism for vertex resolution.
• Utilizing Harsanyi-Farrand-Chang method for wavelength estimation and improved non-negative matrix factorization for separation.

Main Results:

• Successful recovery of multispectral objects from separated speckles across six wavelength channels.
• Validation of the method for imaging hidden objects around corners, improving surround view capabilities.
• Demonstrated high-quality multispectral imaging performance in challenging scattering environments.

Conclusions:

• The proposed method significantly advances multispectral imaging in scattering media.
• This technique offers a promising solution for enhanced spectral recovery and imaging around corners.
• The approach has broad applicability for various scientific and industrial applications requiring advanced imaging.