Robust Texture Mapping Using RGB-D Cameras

Area of Science:

• Computer Vision
• 3D Reconstruction
• Computer Graphics

Background:

• Textured 3D mesh generation from RGB-D images often suffers from visual artifacts.
• Inaccurate camera pose estimation, leading to misalignments, is a primary cause of these artifacts.
• Accumulated errors in pose estimation pose a significant challenge for traditional methods.

Purpose of the Study:

• To develop a robust texture mapping methodology that overcomes camera pose estimation inaccuracies.
• To improve the quality of textured 3D meshes generated from RGB-D data, even with considerable misalignments.
• To leverage depth data from RGB-D images to enhance texture mapping robustness.

Main Methods:

• Proposing a novel texture mapping procedure that utilizes depth information from RGB-D images.
• Developing a method to compensate for non-neglectable errors in camera pose estimations.
• Integrating depth data to enhance the alignment and blending of textures.

Main Results:

• The proposed texture mapping procedure significantly improves the quality of textured 3D meshes.
• The method demonstrates robustness in scenarios with considerable camera misalignments.
• Visual artifacts in textured 3D meshes are substantially reduced.

Conclusions:

• Camera pose estimation errors are inherent and require robust texture mapping solutions.
• Utilizing depth data from RGB-D images is an effective strategy for robust texture mapping.
• The developed method offers a significant advancement in generating high-quality textured 3D meshes from imperfect data.