首页分享基于下采样的自监督点云去噪方法

基于下采样的自监督点云去噪方法

来源：花匠小妙招时间：2025-02-08 10:16

摘要： 针对无噪点云采集困难且使用模拟噪声在合成数据集上训练泛化性能较低的问题, 提出一种仅需含噪点云即可完成训练的自监督去噪方法, 以实现在不同环境下的点云去噪. 该方法首先通过设计并实现特殊的采样器, 对带噪点云下采样, 得到训练网络所需的配对点云；然后通过设计轻型多尺度去噪网络, 解决网络训练中噪声的扰动问题. 在多个数据集上的实验结果表明该方法有效, 在不同场景下能获得与有监督训练相同的效果.

关键词: 自监督学习, 点云去噪, 下采样, 深度学习

Abstract: Aiming at the problem of the difficulty in collecting noiseless point clouds and the low generalisation performance of training on synthetic datasets using simulated noise, we proposed a self-supervised denoising method that only required noisy point clouds to complete training in order to achieve point cloud denoising in different environments. The method first performed downsampling on the noisy point cloud by designing and implementing a special sampler to obtain the paired point cloud required for training the network, and then the problem of noise perturbation in network training was solved by designing a lightweight multi-scale denoising network. The experimental results on multiple datasets show that the method is effective and can obtain the same effect as supervised training in different scenarios.

Key words: self-supervised learning, point cloud denoising, downsampling, deep learning

中图分类号:

TP391.41

引用本文

侯广哲, 秦贵和, 梁艳花. 基于下采样的自监督点云去噪方法[J]. 吉林大学学报(理学版), 2024, 62(1): 100-0105.

HOU Guangzhe, QIN Guihe, LIANG Yanhua. Self-supervised Point Cloud Denoising Method Based on Downsampling[J]. Journal of Jilin University Science Edition, 2024, 62(1): 100-0105.

使用本文

/ 收藏文章 0 / 推荐

导出引用管理器 EndNote|Reference Manager|ProCite|BibTeX|RefWorks

链接本文:http://xuebao.jlu.edu.cn/lxb/CN/

http://xuebao.jlu.edu.cn/lxb/CN/Y2024/V62/I1/100