D-Net: A Density-Based Convolutional Neural Network for Mobile LiDAR Point Clouds Classification in Urban Areas
The 3D semantic segmentation of a LiDAR point cloud is essential for various complex infrastructure analyses such as roadway monitoring, digital twin, or even smart city development. Different geometric and radiometric descriptors or diverse combinations of point descriptors can extract objects from...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-04-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/15/9/2317 |
| _version_ | 1797601803447042048 |
|---|---|
| author | Mahdiye Zaboli Heidar Rastiveis Benyamin Hosseiny Danesh Shokri Wayne A. Sarasua Saeid Homayouni |
| author_facet | Mahdiye Zaboli Heidar Rastiveis Benyamin Hosseiny Danesh Shokri Wayne A. Sarasua Saeid Homayouni |
| author_sort | Mahdiye Zaboli |
| collection | DOAJ |
| description | The 3D semantic segmentation of a LiDAR point cloud is essential for various complex infrastructure analyses such as roadway monitoring, digital twin, or even smart city development. Different geometric and radiometric descriptors or diverse combinations of point descriptors can extract objects from LiDAR data through classification. However, the irregular structure of the point cloud is a typical descriptor learning problem—how to consider each point and its surroundings in an appropriate structure for descriptor extraction? In recent years, convolutional neural networks (CNNs) have received much attention for automatic segmentation and classification. Previous studies demonstrated deep learning models’ high potential and robust performance for classifying complicated point clouds and permutation invariance. Nevertheless, such algorithms still extract descriptors from independent points without investigating the deep descriptor relationship between the center point and its neighbors. This paper proposes a robust and efficient CNN-based framework named D-Net for automatically classifying a mobile laser scanning (MLS) point cloud in urban areas. Initially, the point cloud is converted into a regular voxelized structure during a preprocessing step. This helps to overcome the challenge of irregularity and inhomogeneity. A density value is assigned to each voxel that describes the point distribution within the voxel’s location. Then, by training the designed CNN classifier, each point will receive the label of its corresponding voxel. The performance of the proposed D-Net method was tested using a point cloud dataset in an urban area. Our results demonstrated a relatively high level of performance with an overall accuracy (OA) of about 98% and precision, recall, and F1 scores of over 92%. |
| first_indexed | 2024-03-11T04:07:52Z |
| format | Article |
| id | doaj.art-5aef3f603d2c4188a951ab2e91af9368 |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-11T04:07:52Z |
| publishDate | 2023-04-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-5aef3f603d2c4188a951ab2e91af93682023-11-17T23:38:29ZengMDPI AGRemote Sensing2072-42922023-04-01159231710.3390/rs15092317D-Net: A Density-Based Convolutional Neural Network for Mobile LiDAR Point Clouds Classification in Urban AreasMahdiye Zaboli0Heidar Rastiveis1Benyamin Hosseiny2Danesh Shokri3Wayne A. Sarasua4Saeid Homayouni5Department of Photogrammetry and Remote Sensing, School of Surveying and Geospatial Engineering, College of Engineering, University of Tehran, Tehran 141663, IranDepartment of Photogrammetry and Remote Sensing, School of Surveying and Geospatial Engineering, College of Engineering, University of Tehran, Tehran 141663, IranDepartment of Photogrammetry and Remote Sensing, School of Surveying and Geospatial Engineering, College of Engineering, University of Tehran, Tehran 141663, IranDepartment of Photogrammetry and Remote Sensing, School of Surveying and Geospatial Engineering, College of Engineering, University of Tehran, Tehran 141663, IranGlenn Department of Civil Engineering, Clemson University, Clemson, SC 29634, USACentre Eau Terre Environnement, Institut National de la Recherche Scientifique, 490 Rue de la Couronne, Quebec City, QC G1K 9A9, CanadaThe 3D semantic segmentation of a LiDAR point cloud is essential for various complex infrastructure analyses such as roadway monitoring, digital twin, or even smart city development. Different geometric and radiometric descriptors or diverse combinations of point descriptors can extract objects from LiDAR data through classification. However, the irregular structure of the point cloud is a typical descriptor learning problem—how to consider each point and its surroundings in an appropriate structure for descriptor extraction? In recent years, convolutional neural networks (CNNs) have received much attention for automatic segmentation and classification. Previous studies demonstrated deep learning models’ high potential and robust performance for classifying complicated point clouds and permutation invariance. Nevertheless, such algorithms still extract descriptors from independent points without investigating the deep descriptor relationship between the center point and its neighbors. This paper proposes a robust and efficient CNN-based framework named D-Net for automatically classifying a mobile laser scanning (MLS) point cloud in urban areas. Initially, the point cloud is converted into a regular voxelized structure during a preprocessing step. This helps to overcome the challenge of irregularity and inhomogeneity. A density value is assigned to each voxel that describes the point distribution within the voxel’s location. Then, by training the designed CNN classifier, each point will receive the label of its corresponding voxel. The performance of the proposed D-Net method was tested using a point cloud dataset in an urban area. Our results demonstrated a relatively high level of performance with an overall accuracy (OA) of about 98% and precision, recall, and F1 scores of over 92%.https://www.mdpi.com/2072-4292/15/9/2317point cloud classificationdeep learningvoxelizationautomated object detectionmobile laser scanning |
| spellingShingle | Mahdiye Zaboli Heidar Rastiveis Benyamin Hosseiny Danesh Shokri Wayne A. Sarasua Saeid Homayouni D-Net: A Density-Based Convolutional Neural Network for Mobile LiDAR Point Clouds Classification in Urban Areas Remote Sensing point cloud classification deep learning voxelization automated object detection mobile laser scanning |
| title | D-Net: A Density-Based Convolutional Neural Network for Mobile LiDAR Point Clouds Classification in Urban Areas |
| title_full | D-Net: A Density-Based Convolutional Neural Network for Mobile LiDAR Point Clouds Classification in Urban Areas |
| title_fullStr | D-Net: A Density-Based Convolutional Neural Network for Mobile LiDAR Point Clouds Classification in Urban Areas |
| title_full_unstemmed | D-Net: A Density-Based Convolutional Neural Network for Mobile LiDAR Point Clouds Classification in Urban Areas |
| title_short | D-Net: A Density-Based Convolutional Neural Network for Mobile LiDAR Point Clouds Classification in Urban Areas |
| title_sort | d net a density based convolutional neural network for mobile lidar point clouds classification in urban areas |
| topic | point cloud classification deep learning voxelization automated object detection mobile laser scanning |
| url | https://www.mdpi.com/2072-4292/15/9/2317 |
| work_keys_str_mv | AT mahdiyezaboli dnetadensitybasedconvolutionalneuralnetworkformobilelidarpointcloudsclassificationinurbanareas AT heidarrastiveis dnetadensitybasedconvolutionalneuralnetworkformobilelidarpointcloudsclassificationinurbanareas AT benyaminhosseiny dnetadensitybasedconvolutionalneuralnetworkformobilelidarpointcloudsclassificationinurbanareas AT daneshshokri dnetadensitybasedconvolutionalneuralnetworkformobilelidarpointcloudsclassificationinurbanareas AT wayneasarasua dnetadensitybasedconvolutionalneuralnetworkformobilelidarpointcloudsclassificationinurbanareas AT saeidhomayouni dnetadensitybasedconvolutionalneuralnetworkformobilelidarpointcloudsclassificationinurbanareas |