Preprocessed Data of OSDaR23 Dataset for Publication "Metric Scaling and Extrinsic Calibration of Monocular Neural Network-Derived 3D Point Clouds in Railway Applications"

These are the preprocessed data from the OSDaR23 dataset (https://data.fid-move.de/dataset/osdar23), provided to support the reproduction of results from the publication "Metric Scaling and Extrinsic Calibration of Monocular Neural Network-Derived 3D Point Clouds in Railway Applications" (https://www.mdpi.com/2076-3417/15/10/5361). The file includes: - Preprocessed segmentations generated using InternImage with our pretrained model (https://github.com/OpenGVLab/InternImage/tree/master/segmentation) - Depth images generated using UniDepth v2 (https://github.com/lpiccinelli-eth/UniDepth) To reproduce the results, please follow the instructions on our GitHub page: https://github.com/dthomane/DispImgScaleCalib. Many thanks to: - Roman Tilly , Philipp Neumaier , Karsten Schwalbe , Pavel Klasek , Rustam Tagiew , Patrick Denzler , Tobias Klockau , Martin Boekhoff , Martin Köppel , (2023). Open Sensor Data for Rail 2023 [Data set]. TIB. https://doi.org/10.57806/9mv146r0 - Wang, Wenhai & Dai, Jifeng & Chen, Zhe & Zhenhang, Huang & Li, Zhiqi & Zhu, Xizhou & Hu, Xiaowei & Lu, Tong & Lu, Lewei & Li, Hongsheng & Wang, Xiaogang & Qiao, Yu. (2022). InternImage: Exploring Large-Scale Vision Foundation Models with Deformable Convolutions. 10.48550/arXiv.2211.05778. - Piccinelli, Luigi & Yang, Yung-Hsu & Sakaridis, Christos & Segu, Mattia & Li, Siyuan & Gool, Luc & Yu, Fisher. (2024). UniDepth: Universal Monocular Metric Depth Estimation. 10106-10116. 10.1109/CVPR52733.2024.00963.