Frame by Frame Analysis for Counting of Vehicles from Aerial Video

Surveillance of roads using Unmanned Aerial Vehicles (UAV's) is an upcoming technology which has advantages over existing automatic traffic control methods. Irrespective of the fact that the existing video analysis gives 85-90 percent accuracy, they have a few faults in calculating the exact count of vehicles. So, our project involves frame by frame analysis of videos for better results. The framework includes three steps which are detection of foreground, detection of blobs, dilation and counting of vehicles.