Many large cities in Myanmar are suffering from severe traffic problems as their vehicle populations have been being unprecedentedly increase. Nowadays, due to busy schedule of work style, the people are facing heavy traffic problem. This research presents the automatic vehicles counting based on the density of vehicles in a particular road using image processing techniques. In this work, video image processing to extract video frame as input & using detector. The surveillance cameras are fixed at the top view on that road to capture all vehicle motions. Then, frame images are retrieved from the live video and vehicle regions are segmented applying Gaussian Mixture Model and Morphological operations such as opening and closing. After that, motion vehicle regions are appeared as foreground regions and counted. The accuracy of the proposed method will presents with various experiments of different roads.