A Rail Profile Measurement Method Based on Polarization Fusion Imaging

Area of Science:

• Optical Engineering
• Metrology
• Materials Science

Background:

• Smooth rail surfaces cause underexposure in laser section images, leading to rail profile measurement errors.
• Traditional measurement techniques struggle with data acquisition in these challenging regions.

Purpose of the Study:

• To develop a novel rail profile measurement technique using polarization fusion imaging.
• To enhance the accuracy and reliability of rail profile measurements, especially under complex working conditions.

Main Methods:

• Utilized a polarized camera to capture four-directional polarization component images, Stokes parameter images, linear polarization angle images, and linear polarization degree images.
• Developed a polarization image data fusion algorithm based on Segmented Random Sample Consensus (S-RANSAC).

Main Results:

• The proposed method successfully acquired accurate rail profile data in areas where traditional methods failed.
• Measurement error was reduced from 0.137 mm to 0.081 mm, an improvement of 40.9% in accuracy.
• The technique effectively mitigates data loss caused by local underexposure.

Conclusions:

• Polarization fusion imaging offers a robust solution for high-precision rail profile measurement.
• This method significantly enhances measurement accuracy and reliability under challenging conditions.
• The technique provides a valuable reference for advanced rail inspection and maintenance.