LiDAR Data Classification Using Spatial Transformation and CNN
LiDAR Data Classification Using Spatial Transformation and CNN
复制标题
使用空间变换和 CNN 进行 LiDAR 数据分类
DOI：
10.1109/lgrs.2018.2868378
复制
发表时间：
2019-01-01
影响因子：
4.8
通讯作者：
Chen, Yushi
中科院分区：
文献类型：
--
作者：
He, Xin;Wang, Aili;Chen, Yushi
光探测与测距（LiDAR）是一种有用的数据采集技术，广泛应用于各种实际应用中。激光雷达栅格化数字表面模型分类是激光雷达数据处理的基础技术之一。近年来，深度学习方法，特别是卷积神经网络（CNN），已经在遥感领域显示出了它们的能力，包括LiDAR数据处理。传统的深度模型根据经验使用固定的邻域系统作为网络的输入。因此，输入矩形的权重和高度可能不是最佳的。为了修改这种手工制作的设置，这里使用空间变换网络来识别最佳输入。转换后的输入被送入一个精心设计的CNN，以获得最终的分类结果。此外，形态轮廓与空间变换CNN相结合，以进一步提高分类精度。所提出的框架在两个LiDAR-DSM上进行了测试（即，Recology和Houston数据集）。实验结果表明，与最先进的方法相比，所提出的模型提供了有竞争力的结果。此外，所提出的最佳输入识别方法也可以发现有益的其他遥感应用。
Light detection and ranging (LiDAR) is a useful data acquisition technique, which is widely used in a variety of practical applications. The classification of LiDAR-derived rasterized digital surface model (LiDAR-DSM) is a fundamental technique in LiDAR data processing. In recent years, deep learning methods, especially convolutional neural networks (CNNs), have shown their capability in remote sensing areas, including LiDAR data processing. Traditional deep models empirically use a fixed neighborhood system as input to the network. Therefore, the weight and height of the input rectangle may not be optimal. In order to modify such handcrafted setting, a spatial transformation network is used here to identify optimal inputs. The transformed inputs are fed into a well-designed CNN to obtain the final classification results. Furthermore, morphological profiles are combined with spatial transformation CNN to further improve the classification accuracy. The proposed frameworks are tested on two LiDAR-DSMs (i.e., the Recology and Houston data sets). The experimental results show that the proposed models provide competitive results compared to the state-of-the-art methods. Furthermore, the proposed optimal input identification approach can also be found beneficial for other remote sensing applications.