Research on a hole filling algorithm of a point cloud based on structure from motion.

Area of Science:

• 3D Computer Vision
• Computational Geometry
• Metrology

Background:

• Point clouds from fringe projection often contain holes, hindering 3D reconstruction.
• These holes arise from complex object geometries and inherent method limitations.
• Accurate 3D data processing relies on complete point cloud information.

Purpose of the Study:

• To develop and validate a robust method for filling holes in 3D point clouds.
• To enhance the completeness and integrity of point cloud data for downstream applications.
• To restore surface information in areas with complex shapes.

Main Methods:

• Hole boundary extraction using 2D phase from fringe projection.
• Registration between Structure from Motion (SFM) and fringe projection point clouds.
• Supplementary point extraction and hole filling via radial basis function interpolation.

Main Results:

• Successful identification and filling of holes in complex 3D point clouds.
• Demonstrated robustness of the proposed hole-filling technique through experimental validation.
• Sufficient restoration of surface information in previously incomplete data.

Conclusions:

• The proposed SFM-based fitting approach effectively addresses point cloud hole-filling challenges.
• This method significantly improves the quality of 3D reconstructed data.
• The technique offers a viable solution for handling complex surfaces in 3D measurement.