The 3-D point cloud is a common 3-D data representation that has received increasing attention for remote sensing applications. However, processing 3-D point cloud semantics, especially local semantic information, has always been a challenge and has attracted much attention. In this article, we propose a novel enhanced local semantic learning transformer for 3-D point cloud analysis, which aims to enhance the transformer awareness of local semantic features to handle complex point cloud tasks. First, we propose a novel transformer framework, the local semantic learning point cloud transformer (LSLPCT), which not only learns 3-D point clouds of global information, but also enhances the perception of local semantic information end-to-end. Second, we design an efficient local semantic learning self-attention mechanism, namely, LSL-SA, which can parallelize the perception of global contextual information and capture finer grained local semantic features. Third, our proposed LSL-SA is easy to implement and can integrate the existing transformers and convolutional neural network (CNN)-based networks for processing various point cloud tasks. Numerous experiments in different types of point cloud tasks have been conducted, and our method performs better or is competitive with other state-of-the-art methods.

中文翻译：

---

LSLPCT：用于 3-D 点云分析的增强型局部语义学习转换器

3-D 点云是一种常见的 3-D 数据表示，在遥感应用中越来越受到关注。然而，处理 3D 点云语义，尤其是局部语义信息，一直是一个挑战，备受关注。在本文中，我们提出了一种用于 3-D 点云分析的新型增强型局部语义学习转换器，旨在增强转换器对局部语义特征的认识，以处理复杂的点云任务。首先，我们提出了一种新颖的变换器框架，局部语义学习点云变换器（LSLPCT），它不仅可以学习全局信息的 3D 点云，还可以端到端地增强对局部语义信息的感知。其次，我们设计了一种高效的局部语义学习自注意力机制，即 LSL-SA，它可以并行化对全局上下文信息的感知并捕获更细粒度的局部语义特征。第三，我们提出的 LSL-SA 易于实现，并且可以集成现有的变压器和基于卷积神经网络 (CNN) 的网络来处理各种点云任务。已经在不同类型的点云任务中进行了大量实验，我们的方法表现更好，或者与其他最先进的方法相比具有竞争力。