| 摘要: |
| 现有基于深度学习的点云配准方法主要聚焦于特征提取和特征匹配方面的研究, 然而, 其在特征提取阶段对局部和全局图结构的挖掘尚不充分, 同时在匹配过程中对差异信息的探索也较为有限. 为此, 提出了一种局部-全局动态图学习与互补融合的点云配准方法. 具体而言, 动态偏移的局部图学习模块通过构造包含几何和语义信息的代理点来刻画特征空间中潜在的图结构, 从而获得更具判别性的局部特征. 其次, 设计了动态关注的全局图学习模块, 根据点之间的相互关系自适应地调整关注权重, 有效地捕获了点云中的长程依赖关系. 为了进一步提高两个点云之间的对应关系, 构造了注意力驱动的互补融合模块, 根据交叉注意力机制来挖掘相似信息和差异信息, 并利用自注意力机制优化特征之间的关联性. 实验结果表明, 该方法在公开数据集上实现了最优的配准效果, 并具备良好的计算效率. |
| 关键词: 深度学习 点云配准 图神经网络 特征提取 特征匹配 |
| DOI：10.13328/j.cnki.jos.007416 |
| 分类号:TP181 |
| 基金项目:国家自然科学基金(62032022, 62176244) |
|
| Point Cloud Registration Method Based on Local-global Dynamic Graph Learning and Complementary Fusion |
|
QIU Qiao-Yan1, YE Hai-Liang1, CAO Fei-Long1, LYU Ke2
|
|
1.College of Sciences, China Jiliang University, Hangzhou 310018, China;2.School of Engineering Science, University of Chinese Academy of Sciences, Beijing 100049, China
|
| Abstract: |
| Existing deep learning-based point cloud registration methods primarily focus on feature extraction and feature matching. However, the exploration of local and global graph structures during the feature extraction stage remains insufficient, and the investigation of difference information during the matching process is also limited. To address these issues, this study proposes a point cloud registration method based on local-global dynamic graph learning and complementary fusion. Specifically, the dynamic offset-based local graph learning module characterizes the underlying graph structure in the feature space by constructing proxy points that contain both geometric and semantic information, leading to more discriminative local features. In addition, a dynamic attention-based global graph learning module is designed, which adaptively adjusts attention weights based on the relationships between points, effectively capturing long-range dependencies in the point cloud. To further enhance the correspondence between the two point clouds, the attention-driven complementary fusion module utilizes the cross-attention mechanism to extract similar and distinctive information, while applying the self-attention mechanism to refine the relationships between features. Experimental results demonstrate that the proposed method achieves optimal registration performance on public datasets while maintaining acceptable computational efficiency. |
| Key words: deep learning point cloud registration graph neural network (GNN) feature extraction feature matching |
