Compressed ultrahigh-speed single-pixel imaging by swept aggregate patterns.

Area of Science:

• Optics and Photonics
• Computational Imaging

Background:

• Single-pixel imaging (SPI) offers a unique approach to image acquisition using coded illumination and a single-point detector.
• Existing SPI systems face limitations due to slow digital micromirror device (DMD) refresh rates and computationally intensive compressed-sensing (CS) reconstruction.
• Prior attempts to increase SPI speed, like mechanical masks, compromised reconfigurability and accuracy.

Purpose of the Study:

• To develop an accelerated single-pixel imaging system (SPI-ASAP) that overcomes the speed and reconfigurability limitations of conventional SPI.
• To introduce a novel, lightweight compressed-sensing (CS) reconstruction algorithm optimized for high-speed video streaming.

Main Methods:

• SPI-ASAP integrates a digital micromirror device (DMD) with laser scanning hardware for rapid pattern projection (up to 14.1 MHz).
• A new CS reconstruction algorithm, based on S-cyclic matrices, is developed for efficient, parallelized computation.
• The system supports tunable frame sizes up to 101x103 pixels.

Main Results:

• Achieved pattern projection rates up to 14.1 MHz, significantly enhancing imaging speed.
• Developed a real-time video streaming capability at 100 frames per second (fps) using the lightweight CS algorithm.
• Demonstrated reconfigurable imaging in transmission and reflection modes, dynamic imaging in bright ambient light, and offline ultrahigh-speed imaging up to 12,000 fps.

Conclusions:

• SPI-ASAP represents a significant advancement in single-pixel imaging technology, offering unprecedented speed and flexibility.
• The developed system addresses key bottlenecks in SPI, enabling new applications in dynamic and high-speed imaging.
• The combination of hardware acceleration and efficient algorithms paves the way for real-time, high-performance single-pixel imaging.