Computational imaging with spectral coding increases the spatial resolution of fiber optic bundles

Area of Science:

• Optics
• Biomedical Engineering
• Imaging Science

Background:

• Fiber optic bundles are crucial for imaging in confined spaces but have limited spatial resolution.
• Current methods capture only one pixel per fiber core, neglecting cladding information, thus restricting image detail.

Purpose of the Study:

• To overcome the fundamental spatial resolution limitations of fiber optic bundle imaging.
• To improve image fidelity by extracting more information per fiber and utilizing inter-fiber cladding data.

Main Methods:

• Developed a computational imaging (CI) approach integrated with spectral coding.
• Utilized high-resolution mask patterns to acquire multiple scene images, enabling recovery of up to 17 pixels per fiber core.
• Incorporated a dispersive element to spectrally shift light, allowing transmission of previously lost cladding information through adjacent cores.

Main Results:

• Successfully resolved object features 5x smaller than individual fiber cores.
• Demonstrated significant improvement over conventional imaging, which could only resolve features 1.5x larger than fiber cores.
• Experimental validation performed using both synthetic and real objects.

Conclusions:

• CI combined with spectral coding effectively overcomes the two primary limitations of fiber optic bundle imaging.
• This novel approach substantially enhances spatial resolution, enabling more detailed imaging in medical and industrial applications.
• The technique offers a pathway to significantly improved performance for fiber-based imaging systems.