In this research, a new Urban Point Cloud Dataset obtained by Mobile Laser Scanning (MLS) for Automatic Segmentation and Classification is presented. We go over the steps involved in obtaining the dataset, from acquisition to labelling and post-processing. This dataset can be used to train pointwise classification algorithms, but it can also be used to train object recognition and segmentation algorithms because of the careful attention that has been paid to the divide between the various objects. About 2 km of MLS point cloud that were collected in two cities make up the dataset. The quantity of points and variety of classes lead us to believe that Deep-Learning techniques can be trained with it. Additionally, we display a few outcomes of our automated segmentation and categorization. The dataset is available at: http://caor-mines-paristech.fr/fr/paris-lille-3d-dataset/.

Urban Point Cloud, Dataset, Classification, Segmentation, Mobile Laser Scanning