SeDAR Dataset

Dataset captured as part of our IJCV Publication on Semantic Localisation.

The datset consists of 5 sequences [0-5] caputed on a Turtlebot travelling around CVSSP's offices. Sequence 0 is used for training in the paper, while sequences 1-5 are used for testing. The sequences are captured at different times, meaning they contain variability in the visual domain. These trajectories include static and dynamic obstacles, people, changing geometry and other difficulties. Paper PDF

  • Whole Dataset
  • Training Set
  • Test Set: 1, 2, 3, 4, 5
  • Image Extraction
  • Maps
  • License.txt
  • Contact Me: o.mendez@surrey.ac.uk
    • Researcher in Robotics, Computer Vision, Machine Learning and AI.