Abstract

Dynamic weather effects such as rain cause rapid, distracting motion in a video sequence. This paper aims to remove rain and similar effects from video footage using a multi-step approach; Regions are identified as being potentially affected by rain if they exhibit a short-duration intensity spike. Falling rain drops are imaged by a video camera in a predictable way, as a streak with a consistent range of possible aspect ratios. To preserve scene motion, regions identified by this criterion are investigated, and those that do not fit into the expected range of aspect ratios are ignored. Information about the direction of rainfall is also used to reduce false detections. The effectiveness of this technique is shown on a number of video sequences. The method presented provides advantages over existing techniques.

References

  1. 1.
    Hase, H., Miyake, K., Yoneda, M.: Real-time snowfall noise elimination. In: Proceedings of 1999 International Conference on Image Processing, ICIP 1999, vol. 2, pp. 406–409 (1999)Google Scholar
  2. 2.
    Starik, S., Werman, M.: Simulation of rain in videos. In: Texture Workshop, ICCV, vol. 2, pp. 406–409 (2003)Google Scholar
  3. 3.
    Barnum, P., Kanade, T., Narasimhan, S.: Spatio-temporal frequency analysis for removing rain and snow from videos. In: Workshop on Photometric Analysis For Computer Vision (2007)Google Scholar
  4. 4.
    Garg, K., Nayar, S.: Vision and rain. International Journal of Computer Vision 75, 3–27 (2007)CrossRefGoogle Scholar
  5. 5.
    Zhang, X., Li, H., Qi, Y., Leow, W.K., Ng, T.K.: Rain removal in video by combining temporal and chromatic properties. In: IEEE International Conference on Multimedia and Expo 2006, pp. 461–464 (July 2006)Google Scholar
  6. 6.
    Foote, G.B., Du Toit, P.S.: Terminal Velocity of Raindrops Aloft. Journal of Applied Meteorology 8, 249–253 (1969)CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2008

Authors and Affiliations

  • Nathan Brewer
    • 1
    • 2
  • Nianjun Liu
    • 1
    • 2
  1. 1.RSISE, The Australian National UniversityAustralia
  2. 2.NICTA Canberra LabAustralia

Personalised recommendations