Specular highlight suppression based on full polarization imaging.

Area of Science:

• Computer Vision
• Optical Engineering
• Materials Science

Background:

• Specular highlights pose significant challenges in industrial inspection, obscuring critical features on reflective surfaces.
• Current specular highlight suppression techniques are inadequate for materials with high reflectivity.

Purpose of the Study:

• To develop an advanced specular highlight suppression method utilizing full polarization imaging.
• To enhance feature extraction accuracy in industrial inspection scenarios involving highly reflective materials.

Main Methods:

• Constructed a full polarization image dataset using the quarter-waveplate rotation method.
• Derived an analytical model for full polarization feature parameters.
• Developed an adaptive specular highlight detection method using polarization feature clusters and region expansion.
• Implemented an edge curve integration method for compensation value calculation to improve image stitching.

Main Results:

• The proposed method effectively suppresses specular highlights on various surfaces.
• Experimental results on self-built (FPID) and public datasets show superior visual effects and quantitative metrics compared to existing approaches.
• The method successfully handles surfaces with high reflective characteristics.

Conclusions:

• The full polarization imaging-based method offers a superior solution for specular highlight suppression in industrial inspection.
• This technique significantly enhances the reliability of feature extraction on challenging reflective surfaces.
• The developed adaptive detection and compensation methods contribute to robust performance across diverse datasets.