Real-Time Pedestrian Recognition in Urban Environments

  • Basam Musleh
  • Arturo de la Escalera
  • José Maria Armingol
Part of the VDI-Buch book series (VDI-BUCH)


Traditionally, pedestrian recognition is a great research topic in computer vision applied to advanced driver assistance systems ( ADAS); a real-time pedestrian recognition system based on stereo vision is presented in this paper. The most interesting features of the system are that it does not need any extrinsic calibration and it is possible to determine the pedestrian’s localization with a bigger resolution than only by using the disparity values. This is possible because the road profile in front of the vehicle is calculated from the v-disparity at each frame. Once the road profile has been generated the obstacles can be classified into elevated obstacles or obstacles on the ground. Regarding the pedestrian recognition, a fast method has been developed based on the similarity between the vertical projection of the pedestrian’s silhouette and a normal distribution. Stereo algorithms have a high computation time and in order to cope with this, our algorithm has been implemented in graphics processing unit by means of CUDA.


Stereo Vision Vertical Projection Obstacle Detection Advanced Driver Assistance System Road Profile 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. [1]
    Scharstein, D, Szeliski, R, A taxonomy and evaluation of dense two-frame stereo correspondence algorithms, International Journal of Computer Vision, 47, 7-42, 2002.zbMATHCrossRefGoogle Scholar
  2. [2]
    Soquet, N, Perrollaz, M, Labayrade, R, Auber, D, Free space estimation for autonomous navigation, Proceedings of the 5th Int. Conference on Comput. Vision Systems, 1-6, 2007.Google Scholar
  3. [3]
    Broggi, A, Caraffi, C, Fedriga, R, I, Grisleri, Obstacle detection with stereo vision for off-road vehicle vehicle navigation, Proceedings of the IEEE Conference on Computer vision and Pattern Recognition, 1-6, 2005.Google Scholar
  4. [4]
    Labayrade, R, Aubert, D, Tarel, J, P, Real time obstacles detection in stereovision on non flat road geometry through V-disparity representation vehicle navigation, Intelligent Vehicle Symposium. 1-6, 2002.Google Scholar
  5. [5]
    NVIDIA CUDA, Programming guide, 2.3.1 version, NVIDIA Co.Google Scholar
  6. [6]
    Lee, C,H, Lim Y,C, Kong, S, Lee, J,H, Obstacle localization with a binarized v-disparity map using local maximum frequency values in stereo vision. Proceedings of the International Conference on Signals, Circuits and Systems, 1-4, 2008.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2011

Authors and Affiliations

  • Basam Musleh
    • 1
  • Arturo de la Escalera
    • 1
  • José Maria Armingol
    • 1
  1. 1.Escuela Politécnica SuperiorUniversity Carlos III of MadridLeganésSpain

Personalised recommendations