arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2301.13656cs.CVcs.CG
收藏

从点云自动重建表面的综述与基准

A Survey and Benchmark of Automatic Surface Reconstruction from Point Clouds

Centre INRIA d’Université Côte d’Azur(蔚蓝海岸大学法国国家信息与自动化研究所中心);LIGM, Ecole des Ponts, Univ Gustave Eiffel, CNRS(巴黎路桥学校古斯塔夫·埃菲尔大学法国国家科学研究中心LIGM实验室);LASTIG, Univ Gustave Eiffel, IGN-ENSG(古斯塔夫·埃菲尔大学法国地理学院LASTIG实验室)

Raphael Sulzer, Renaud Marlet, Bruno Vallet, Loic Landrieu

AI总结:

本文综述并标准化基准测试点云表面重建的传统与深度学习方法,比较二者在精度、鲁棒性及真实采集异常下的表现,并公开代码与数据集。

摘要:

我们提出了一项针对从点云进行表面重建的传统方法和基于学习方法的全面综述与基准。由于噪声、离群点、非均匀采样和数据缺失等因素,该任务在真实世界采集数据中尤其具有挑战性。传统方法通常通过对输入点云或所得表面施加手工先验来简化问题,这一过程可能需要繁琐的超参数调优。相比之下,深度学习模型能够直接从数据中学习输入点云和目标表面的属性。我们研究了手工先验与学习先验对表面重建技术精度和鲁棒性的影响。我们以标准化方式评估了多种久经检验的方法和当代方法。当在具有相同特征的点云上进行训练和评估时,基于学习的模型始终能比传统方法生成更高质量的表面,即使在涉及新形状类别的场景中也是如此。然而,传统方法对真实世界三维采集中常见的各类异常表现出更强的适应能力。为了研究社区的利益,我们公开了代码和数据集,以推动基于学习的表面重建的进一步改进。相关资源可在 https://github.com/raphaelsulzer/dsr-benchmark 获取。

英文摘要:

We present a comprehensive survey and benchmark of both traditional and learning-based methods for surface reconstruction from point clouds. This task is particularly challenging for real-world acquisitions due to factors such as noise, outliers, non-uniform sampling, and missing data. Traditional approaches often simplify the problem by imposing handcrafted priors on either the input point clouds or the resulting surface, a process that can require tedious hyperparameter tuning. In contrast, deep learning models have the capability to directly learn the properties of input point clouds and desired surfaces from data. We study the influence of handcrafted and learned priors on the precision and robustness of surface reconstruction techniques. We evaluate various time-tested and contemporary methods in a standardized manner. When both trained and evaluated on point clouds with identical characteristics, the learning-based models consistently produce higher-quality surfaces compared to their traditional counterparts -- even in scenarios involving novel shape categories. However, traditional methods demonstrate greater resilience to the diverse anomalies commonly found in real-world 3D acquisitions. For the benefit of the research community, we make our code and datasets available, inviting further enhancements to learning-based surface reconstruction. This can be accessed at https://github.com/raphaelsulzer/dsr-benchmark .

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑