Multi-line structured light stripes clustering based on a custom iterative window

Area of Science:

• Computer Vision
• 3D Reconstruction
• Optical Metrology

Background:

• Multi-line structured light technology is crucial for 3D reconstruction.
• Stripe confusion due to noise and breakage hinders accurate stereo matching.
• Existing methods struggle with robust stripe clustering in challenging conditions.

Purpose of the Study:

• To propose a robust multi-line structured light stripe clustering method.
• To address stripe confusion and mis-correspondence in binocular stereo vision.
• To enhance the accuracy of laser stripe matching.

Main Methods:

• Object segmentation from the carrier plane.
• Clustering of extracted centerlines using a custom iterative window.
• Stripe numbering via barycentric coordinates and sequential consistency constraints.

Main Results:

• Effective clustering of centerline stripes on measured objects.
• Successful resolution of stripe mis-correspondence issues.
• Demonstrated improvement in multi-line structured light stereo matching.

Conclusions:

• The proposed iterative window clustering method enhances stereo matching accuracy.
• This technique effectively overcomes challenges posed by stripe breakage and noise.
• The method provides a reliable solution for 3D reconstruction using multi-line structured light.