Snapshot spatial-temporal compressive imaging.

Area of Science:

• Optics and Photonics
• Computer Vision
• Image Reconstruction

Background:

• High-speed imaging is crucial for capturing dynamic events.
• Traditional methods often require multiple frames or specialized hardware, limiting speed and resolution.
• Compressive imaging offers a potential solution for efficient data acquisition.

Purpose of the Study:

• To develop a compressive imaging system for simultaneous high-speed video reconstruction from two fields-of-view.
• To achieve high spatial resolution in reconstructed videos from a single snapshot.
• To leverage deep learning for efficient and robust image reconstruction.

Main Methods:

• A single digital micromirror device (DMD) was used for high-speed modulation across two scenes.
• A plug-and-play framework incorporating deep denoising priors was utilized for reconstruction.
• The system captures a single 2D spatial measurement to reconstruct two 3D spatio-temporal data cubes.

Main Results:

• Successfully reconstructed two high-speed videos from two fields-of-view using a single captured frame.
• Achieved reconstruction of two spatial-temporal data cubes, each sized 650x650x20.
• Demonstrated that the DMD modulation did not compromise spatial resolution.

Conclusions:

• The developed compressive imaging system enables efficient high-speed video reconstruction from limited measurements.
• The integration of deep denoising priors significantly enhances reconstruction quality.
• This approach offers a promising direction for advanced imaging applications requiring high speed and resolution.