Single-pixel tracking of fast-moving object using geometric moment detection.

Area of Science:

• Optics and Photonics
• Computational Imaging
• Robotics and Automation

Background:

• Real-time tracking of fast-moving objects is crucial for numerous applications.
• Existing single-pixel imaging techniques face challenges in achieving high frame rates for dynamic targets.

Purpose of the Study:

• To develop a novel single-pixel imaging method for real-time tracking of fast-moving objects.
• To leverage geometric moments for efficient object centroid reconstruction and tracking.

Main Methods:

• Illuminating a moving object with only 3 geometric moment patterns per frame.
• Collecting intensities with a single-pixel detector to determine zero-order and first-order geometric moments.
• Reconstructing and tracking the object's centroid based on measured geometric moments.

Main Results:

• Achieved root mean squared errors of 5.46 pixels (transverse) and 5.53 pixels (axial) compared to a camera system.
• Successfully tracked a moving magnet at an exceptional frame rate of up to 7400 Hz.
• Demonstrated the feasibility of the geometric moment-based approach for ultrafast object tracking.

Conclusions:

• The proposed single-pixel imaging technique offers a new, efficient method for real-time, high-frame-rate object tracking.
• This approach overcomes limitations of traditional methods for tracking fast-moving targets.
• The technique holds significant potential for ultrafast target tracking applications in various fields.