Abstract
The success achieved by deep learning techniques in image labeling has triggered a growing interest in applying deep learning for three-dimensional point cloud classification. To provide better insights into different deep learning architectures and their applications to ALS point cloud classification, this article presents a comprehensive comparison among three state-of-the-art deep learning networks: PointNet++, SparseCNN, and KPConv, on two different ALS datasets. The performances of these three deep learning networks are compared w.r.t. classification accuracy, computation time, generalization ability as well as the sensitivity to the choices of hyper-parameters. Overall, we observed that PointNet++, SparseCNN, and KPConv all outperform Random Forest on the classification results. Moreover, SparseCNN leads to a slightly better classification result compared to PointNet++ and KPConv, while requiring less computation time and memory. At the same time, it shows a better ability to generalize and is less impacted by the different choices of hyper-parameters.
Highlights
ALS is one of the most important techniques for data collection of real-world scenes
This study presented the classification performance of PointNet++,SparseCNN and KPConv on two very different ALS point clouds and conducted a comparison among PointNet++, SparseCNN and KPConv in several aspects
Due to the demanding memory usage, only limited number of points can be fed into KPConv, which leads to a lack of sufficient contextual information for the features learning in large-scale datasets
Summary
ALS is one of the most important techniques for data collection of real-world scenes. Tremendous progress in the automatic classification of ALS point clouds has been achieved in the community of remote sensing and photogrammetry. Researchers have developed various deep learning architectures for 3D point cloud classification, such as PointNet++ [10] and SparseCNN [11]. Considering the impressive results achieved on indoor point clouds, those deep learning frameworks are further applied on the outdoor point clouds in the field of remote sensing and photogrammetry. The models are mostly trained and evaluated on the ISPRS Vaihingen 3D dataset [14, 15, 17, 18] These benchmark point clouds provide the possibility for performance comparison of different models, it only covers a small area with limited diversity in the scenes [12]. Unlike normal comparisons of approaches that are conducted on only one benchmark, we performed a more
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callMore From: IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
