Compressive high-speed stereo imaging.

Area of Science:

• Optics and Photonics
• Computer Vision
• 3D Imaging Technology

Background:

• High-speed 3D imaging is crucial for analyzing dynamic events.
• Traditional stereo imaging systems are limited by sensor frame rates.
• Achieving high frame rates in 3D reconstruction often requires complex or expensive setups.

Purpose of the Study:

• To develop a compressive stereo imaging system capable of reconstructing 3D videos at significantly higher frame rates than conventional methods.
• To enable high-speed 3D scene reconstruction using an asymmetric optical configuration and a novel reconstruction algorithm.
• To provide a robust, inexpensive, and active-illumination-free solution for high-speed 3D imaging.

Main Methods:

• Implementation of an asymmetric stereo imaging system incorporating a high-speed spatial modulator in one camera view.
• Development of a two-step reconstruction algorithm to recover scene irradiance and depth information.
• Utilizing a compressive sensing approach to overcome sensor sampling rate limitations.

Main Results:

• Successful high-speed 3D video reconstruction at 800 frames per second (fps) from 80 fps sensor measurements.
• Demonstration of a 10x increase in effective frame rate compared to the imaging sensor's sampling rate.
• Validation of the system's ability to recover both irradiance and depth information for dynamic scenes.

Conclusions:

• The reported compressive stereo imaging system offers a significant advancement in high-speed 3D video acquisition.
• The asymmetric configuration and reconstruction algorithm effectively achieve high-speed 3D reconstruction.
• This method presents a practical and cost-effective solution for various applications requiring high-speed 3D imaging without active illumination.