Super-resolution time-resolved imaging using computational sensor fusion

Area of Science:

• Optics and Imaging
• Biomedical Engineering
• Photonics

Background:

• High-precision 3D imaging is crucial for LIDAR and FLIM but often requires sacrificing spatial for temporal resolution.
• Acquiring full 3D data cubes rapidly presents significant challenges, necessitating innovative imaging solutions.

Purpose of the Study:

• To develop a sensor fusion method that overcomes the spatial-temporal resolution trade-off in imaging.
• To enhance the spatial resolution of time-resolved imaging techniques using complementary sensor data.

Main Methods:

• A sensor fusion approach combining low-spatial/high-temporal resolution SPAD data with high-spatial/no-temporal resolution CMOS data.
• Image blurring on the SPAD array followed by computational sensor fusion to reconstruct enhanced images.

Main Results:

• Reconstruction of time-resolved images with significantly improved spatial resolution compared to SPAD input.
• Demonstrated numerical upsampling by a factor of [Formula: see text] and experimental upsampling up to [Formula: see text].
• Successful application of the technique in LIDAR and fluorescence lifetime imaging (FLIM) of cancer cells.

Conclusions:

• The developed sensor fusion technique effectively enhances spatial resolution in time-resolved imaging.
• This method enables high spatial resolution SPAD imaging and accelerates FLIM imaging by over an order of magnitude.
• Paves the way for advanced imaging capabilities in LIDAR and microscopy.