Measuring and correcting wobble in large-scale transmission radiography

Area of Science:

• Medical Imaging
• X-ray Technology
• Image Processing

Background:

• Large-scale transmission radiography scanners image vehicles and cargo for threat detection.
• Mechanical instability in these systems causes detector wobble, reducing image precision.
• Accurate image precision is crucial for reliable threat detection.

Purpose of the Study:

• To measure and correct for mechanical wobble in large-scale radiography systems.
• To mitigate the loss of image precision caused by detector array instability.
• To enhance the accuracy of threat detection in scanned cargo.

Main Methods:

• Utilized Beam Position Detectors (BPDs) to measure X-ray beam profiles for wobble estimation.
• Developed a model for image formation with a wobbling detector array.
• Implemented a Random Regression Forest for instantaneous wobble estimation and subsequent image correction.
• Calibrated sensor sensitivities and relative offsets for improved accuracy.

Main Results:

• Successfully corrected for 87% of image errors attributed to wobble.
• Observed significant visible improvements in image quality for challenging scenes.
• Demonstrated enhanced precision in wobble-affected images.

Conclusions:

• The developed wobble correction method significantly improves image precision.
• Enhanced image quality aids in better threat detection and material identification.
• This technique is vital for maintaining high-performance security scanning systems.