Classification of ALS Point Clouds Using End-to-End Deep Learning
Classification of ALS Point Clouds Using End-to-End Deep Learning
复制标题
DOI：
10.1007/s41064-019-00073-0
复制
发表时间：
2019-09
期刊：
影响因子：
--
通讯作者：
L. Winiwarter;G. Mandlburger;S. Schmohl;N. Pfeifer
中科院分区：
文献类型：
--
作者：
L. Winiwarter;G. Mandlburger;S. Schmohl;N. Pfeifer
深度学习是指多层人工神经网络，广泛用于包括计算机视觉在内的许多学科的分类任务。最流行的类型是卷积神经网络 (CNN)，通常应用于 2D 图像数据。然而，CNN 很难适应点云等不规则数据。 另一方面，PointNet 可以利用神经网络根据 nD 空间中一组点的几何分布来推导特征。我们在多个尺度上使用 PointNet 以端到端的方式自动学习局部邻域的表示，该表示针对机载激光扫描 (ALS) 获取的 3D 点云上的语义标记进行了优化。结果与使用手动制作的特征的结果相当，表明这些社区的成功表现。在 ISPRS 3D 语义标签基准测试中，我们实现了 80.6% 的整体准确率，属于中等水平。对更大数据集（即联邦福拉尔贝格州 2011 年 ALS 点云）的调查显示，大规模建成区的总体准确率高达 95.8%。低植被和地面点的分离精度较低，可能是因为对空间中类别分布的假设无效，尤其是在高山地区。我们得出的结论是，端到端系统的方法允许对各种分类问题进行训练，而无需有关邻域特征的专业知识，也可以成功应用于 ALS 点云的基于单点的分类。
Deep learning, referring to artificial neural networks with multiple layers, is widely used for classification tasks in many disciplines including computer vision. The most popular type is the Convolutional Neural Network (CNN), commonly applied to 2D image data. However, CNNs are difficult to adapt to irregular data like point clouds.PointNet, on the other hand, has enabled the derivation of features based on the geometric distribution of a set of points in nD-space utilising a neural network. We use PointNet on multiple scales to automatically learn a representation of local neighbourhoods in an end-to-end fashion, which is optimised for semantic labelling on 3D point clouds acquired by Airborne Laser Scanning (ALS). The results are comparable to those using manually crafted features, suggesting a successful representation of these neighbourhoods. On the ISPRS 3D Semantic Labelling benchmark, we achieve 80.6% overall accuracy, a mid-field result. Investigation on a bigger dataset, namely the 2011 ALS point cloud of the federal state of Vorarlberg, shows overall accuracies of up to 95.8% over large-scale built-up areas. Lower accuracy is achieved for the separation of low vegetation and ground points, presumably because of invalid assumptions about the distribution of classes in space, especially in high alpine regions. We conclude that the method of the end-to-end system, allowing training on a big variety of classification problems without the need for expert knowledge about neighbourhood features can also successfully be applied to single-point-based classification of ALS point clouds.