UAVid dataset is a high-resolution UAV semantic segmentation dataset focusing on street scenes.The dataset consists of 30 sequences (seq1 to seq30), which are captured with 4K high-resolution in oblique views.In total, 300 images have been densely labeled with 8 classes for the semantic labeling task.The 8 classes are Building, Road, Tree, Low vegetation, Moving car, Static car, Human, Background clutter.Task DescriptionThe task for UAVid dataset is to predict per-pixel semantic labelling for the UAV imagery, and results will be evaluated with meanIoU metric. The original video files for each sequence will be provided upon request.Data DescriptionTraining data and validation data: each sequence is provided with images and labels. Images are named according to the frame index (0-based) in the video sequence.Test data: only test images are provided.Data expansionWe have collected another 12 sequences for the dataset. They are from seq31 to seq42, and are distributed to train, val and test sets. We recommend to use all data for training and testing. Otherwise, please only use seq1 to seq30 for training and testing as in the original paper.CopyrightUAVid dataset is copyright by us and published under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 License. This means that you must attribute the work in the manner specified by the authors, you may not use this work for commercial purposes and if you alter, transform, or build upon this work, you may distribute the resulting work only under the same license.CitationWhen using this dataset in your research, please cite:@article{uavid20,Author = {Ye Lyu andGeorge Vosselman andGuisong Xia andAlper Yilmaz andMichael Ying Yang},Title = {UAVid: A Semantic Segmentation Dataset for UAV Imagery},journal = {ISPRS Journal of Photogrammetry and Remote Sensing},year = {2020},}Contacty.lyu@utwente.nlmichael.yang@utwente.nl

Date Submitted: 2020-05-19