Vol. 3, Issue 11, Part F (2017)

Abstract
The large variability of road user appearance in an urban setting makes it very difficult to control the traffic and the crowd. Here a real-time system that detects moving crowd in a video sequence is presented. Crowd detection differs from pedestrian detection in that no individual pedestrian can be properly segmented in the image is assumed. A scheme that looks at the motion patterns of crowd in the spatio-temporal domain and give an efficient implementation that can detect crowd in real-time is proposed. Pedestrain detection from a moving platform (i.e, vehicles) holds the promise of increased safety, both for pedestrains and passengers. For tracking users an on board computer vision is developed. Web cam is used to monitor the road side status and the number of vehicles get counted while it enters into the frame. It gives voice suggestions to the authority. Thus, the accuracy of the system is more and accidents can be avoided before it happens.