DL4Lidar dataset - Manually Labelled Lidar Range-Corrected Signal Images for Atmospheric Semantic Segmentation

This dataset contains the images and annotations (in COCO format) used when training, validating, and evaluating the aerosol-cloud segmentation model for lidar range-corrected signal images presented in Canella-Ortiz et al. (2026). The data were collected from the ALHAMBRA lidar instrument, at the UGR station (Granada, Spain), and from the MULHACEN lidar instrument, also at the UGR station. If you use this dataset or the associated code in your research, please cite the corresponding journal paper: Canella-Ortiz, A., Tabik, S., Fernández-Carvelo, S., Rodríguez-Navarro, O., Alados-Arboledas, L., and del Águila, A., "Generalizable Deep Learning for Aerosol and Cloud Segmentation in Ground-Based and Spaceborne Lidar Observations," IEEE Transactions on Geoscience and Remote Sensing, 2026 (In Press). This dataset is designed to be used alongside the official research article repository. For the source code, training scripts, and inference scripts associated with this work and dataset, please visit: https://github.com/acanort/DL4Lidar.