VLMV Logo

Vehicle Lane Merge Visual Benchmark

5GCAR logo

Data Set: Lane Merge Maneuver of Vehicles

Objective 1: Evaluation of vehicle localization techniques
  • Use video data for 3D vehicle localization, GNSS-RTK for evaluation [1]
Objective 2: Learning of lane merge coordination
  • Use GNSS-RTK and vision-based vehicle positions for learning cooperative maneuvers [3]
Data Set:
  • 85 lane merges performed by human drivers on 7 recording days
  • Temporally synchronized multi-view video streams (four cameras)
  • Accurate camera calibration [1,5]
  • Vehicle positions:
    • GNSS-RTK
    • camera-based tracking and localization [1,4]
  • Presentation at ICPR 2020/21:
    Track 1: Artificial Intelligence, Machine Learning for Pattern Analysis
    14. Jan, 12h-13h, Session PS T1.9
    Poster

  • Supplementary video:
    https://youtu.be/tiPABhZoFEw
Download: 7 sets (day01 .. day07)
  • Video data
    • 4 synchronized video streams, 1824x1536, mjpeg format
    • camera calibration
  • Localization data
    • vehicle localization from GNSS-RTK
    • heading, speed, acceleration from in-vehicle measurements
    • camera-based localization
Image
Version 1.0 - all 85 sets (video, calibration, and localization) ready. log.txt

day01:  3 lane merges
video              calibration
localization_GNSS-RTK

day02:  12 lane merges
video              calibration
localization_GNSS-RTK

day03:  4 lane merges
video              calibration
localization_GNSS-RTK

day04:  16 lane merges
video              calibration
localization_GNSS-RTK

day05:  13 lane merges
video              calibration
localization_GNSS-RTK

day06:  27 lane merges
video              calibration
localization_GNSS-RTK

day07:  10 lane merges
video              calibration
localization_GNSS-RTK

Camera-based localization for all VLMV vehicles [1]:  localization_ICPR2020

VLMV Paper:

[1] K. Cordes and H. Broszio: "Vehicle Lane Merge Visual Benchmark", International Conference on Pattern Recognition (ICPR), IEEE, Jan. 2021, Paper at ieeexplore

Additional References:

[2] 5GCAR final demonstration (Video)

[3] O. Nassef, L. Sequeira, E. Salam, & T. Mahmoodi: "Building a Lane Merge Coordination for Connected Vehicles Using Deep Reinforcement Learning" , IEEE Internet of Things Journal, 2020

[4] K. Cordes, N.Nolte, N. Meine, and H. Broszio: "Accuracy Evaluation of Camera-based Vehicle Localization", International Conference on Connected Vehicles and Expo (ICCVE), IEEE, pp. 1-7, Nov. 2019, paper at ieeexplore

[5] K. Cordes and H. Broszio: "Constrained Multi Camera Calibration for Lane Merge Observation", International Conference on Computer Vision Theory and Applications (VISAPP), SciTePress, pp. 529-536, Feb. 2019, paper preprint (pdf)

[6] K. Cordes, H. Broszio, H Wymeersch, S Saur, F Wen, Nil Garcia, Hyowon Kim: "Radio‐Based Positioning and Video‐Based Positioning",
https://doi.org/10.1002/9781119692676.ch8, "Cellular V2X for Connected Automated Driving", Wiley, April 2021
, Book Chapter at ieeexplore