Intelligent Traffic Monitoring Systems: Vehicles Detection, Tracking, And Counting using Haar Cascade Classifier And Optical Flow

In the automation of traffic monitoring system based on video processing, vehicle detection and tracking is an important rule because it is used for further analysis such as calculating traffic flow. This research aims to implement video processing techniques to detect and track vehicle objects. Detection and tracking stage is done to calculate the number of vehicle objects using video senSor. We use Haar Cascade Classifier method for vehicle detection and Optical Flow method for tracking vehicle in Indonesian highway and for vehicle counting process. We use detection region to localize the detection process, tracking and counting the number of vehicles so all process will be done on this region. In the tracking vehicles stage is utilized by centroid from Haar Cascade Classifier detection results. We compare the area in every vehicle detection result from previous frame and current frame with certain threshold value of area to make sure a vehicle is the same object in each frame. So that, in the detection region vehicle will not be repeatedly calculated in the vehicle counting stage. Based on implementation detection test and tracking test result we obtained quite high accuracy.