Automated Video Surveillance for Retail Store Statistics Generation

  • N. Avinash
  • M. S. Shashi Kumar
  • S. M. Sagar
Conference paper
Part of the Lecture Notes in Electrical Engineering book series (LNEE, volume 221)


Video Surveillance Systems have gained immense popularity in the recent past because of the fact that it can be used in numerous real-world scenario applications. Monitoring the people flow pattern as well as counting them serves as valuable information in many surveillance related applications. In this paper we propose a system that is used for counting the number of people passing through the camera field of view. A single overhead camera is used to get a clear top-view which avoids occlusions. For background subtraction, running Gaussian approach has been used as a preprocessing step, to facilitate the further segmentation and tracking procedures. Connected component analysis is used to group the similar blobs together followed by intensity based correlation for blob matching followed by Kalman tracking. The percentage of blobs that crosses a reference line is recorded. Two counters are incremented depending on the direction of movement of the blobs and the algorithm is able to count the number of people moving up/down the scene.


Running Gaussian approach Kalman tracker Connected component analysis Intensity based correlation 


  1. 1.
    Hayman JA (2003) Computer vision based people tracking for motivating behavior in public spaces. Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, pp 121–125Google Scholar
  2. 2.
    Aik LE, Zianuddin Z (2009) Real-time people counting system using curve analysis method. Int J Comput Electr Eng 1(1):1793–8198Google Scholar
  3. 3.
    Adriano GP, Mendoza SIV, Montinola FNJ, Naval PC (2005) APeC: automated people counting form video. In: Computer vision and machine intelligence group, Department of Computer Science, College of Engineering, University of the Philippines-DilimanGoogle Scholar
  4. 4.
    Chen C-H, Chen T-Y, Wang D-J, Chen T-J (2012) A Cost-effective people counter for a crowd of moving people based on two-stage, segmentation. J Inform Hiding Multimedia Signal Process 3(1):2073–4212Google Scholar
  5. 5.
    de Almeida SS, de Melo VHC, Menotti D An Evaluation of two people counting system using zenithal cameraGoogle Scholar
  6. 6.
    Berg R-E (2007) Real-time people counting system using video cameraGoogle Scholar
  7. 7.
    Lu H, Zhang R, Chen Y-W (2008) Head detection and tracking by mean-shift and Kalman filter. Department of Electronics, Dalian University, DalianGoogle Scholar
  8. 8.
    Lempitsky V, Zisserman A (2010) Learning to count objects in images. In: Proceedings of NIPS, pp 1324–1332Google Scholar
  9. 9.
    Conte D, Foggia P, Percannella G, Tufano F, Vento M (2010) Counting moving people in videos by salient points detection. In: Proceedings of ICPR, 2010Google Scholar
  10. 10.
    Lefloch D (2007) Real-time people counting system using video camera. Department of Computer Science and Media Technology, Gjøvik University College, NorwayGoogle Scholar
  11. 11.
    Hsieh J-W, Fang F-J, Lin G-J, Wang Y-S (2012) Template matching and Monte Carlo Markova chain for people counting under occlusions. In: Proceedings of 18th international conference on advances in multimedia modeling, pp 761–771Google Scholar
  12. 12.
    Zhao T, Nevatia R (2003) Bayesian human segmentation in crowded situations. In: Computer vision and pattern recognition, IEEE, pp 459–466Google Scholar
  13. 13.
    Piccardi M (2004) Background subtraction techniques: a review. In: IEEE international conference on systems, man and cybernetics, 2004Google Scholar
  14. 14.
    Ibrahim MM, Anupama R (2005) Scene adaptive shadow detection algorithm. In: Proceedings of world academy of science, engineering and technology, 2005Google Scholar

Copyright information

© Springer India 2013

Authors and Affiliations

  1. 1.Wittybot TechnologiesBangaloreIndia
  2. 2.Manipal Centre for Information Science (MCIS)ManipalIndia

Personalised recommendations