3D-PCGR: Colored Point Cloud Generation and Reconstruction with Surface and Scale Constraints

Chaofeng Yuan,Jinghui Pan,Zhaoxiang Zhang,Min Qi,Yuelei Xu
DOI: https://doi.org/10.3390/rs16061004
IF: 5
2024-03-13
Remote Sensing
Abstract:In the field of 3D point cloud data, the 3D representation of objects is often affected by factors such as lighting, occlusion, and noise, leading to issues of information loss and incompleteness in the collected point cloud data. Point cloud completion algorithms aim to generate complete object point cloud data using partial or local point cloud data as input. Despite promising results achieved by existing methods, current point cloud completion approaches often lack smooth and structural consistency, resulting in a messy overall structure. To address these shortcomings in point cloud completion, we propose a point cloud generative method based on surface consistency and scale rendering. In addition, to solve the limitation of existing methods that mainly focus on geometric features in 3D point cloud completion and do not make full use of color information, we introduce an object reconstruction method based on texture and geometric features. Extensive experiments demonstrate that our proposed methods exhibit superior performance in terms of local details and overall object structure.
environmental sciences,imaging science & photographic technology,remote sensing,geosciences, multidisciplinary
What problem does this paper attempt to address?
The paper attempts to address the issue of information loss and incompleteness in 3D point cloud data caused by factors such as lighting, occlusion, and noise. Although existing point cloud completion methods have achieved certain results, they often lack smoothness and structural consistency, leading to a rather chaotic overall structure. Additionally, existing methods mainly focus on geometric features and fail to fully utilize color information. To overcome these shortcomings, the paper proposes a point cloud generation method based on surface consistency and scale rendering, and introduces an object reconstruction method based on texture and geometric features. Through extensive experimental validation, this method demonstrates excellent performance in both local details and overall object structure. Specifically, the main contributions of the paper include: 1. Proposing a new 3D computer vision algorithm for point cloud completion and object reconstruction, addressing the issues of object structure and information loss caused by noise and occlusion. 2. Designing a point cloud object completion algorithm based on surface consistency and scale rendering. By obtaining surface consistency information through surface consistency constraints, it addresses the issues of insufficient surface details, roughness, high noise, and lack of smoothness in existing point cloud completion algorithms. Utilizing depth and scale information rendering to obtain structural consistency constraints, it improves the surface detail smoothness and overall structural integrity of point cloud completion. 3. Introducing a 3D object point cloud generation method based on pixel-level fusion of color and geometric information. By fully combining color and geometric information through pixel-level fusion, it enhances the completeness and accuracy of object reconstruction.