Automated Vehicle Counting And Classification System

This paper describes a system that can detect, count and classify vehicles automatically based on their types. Some techniques have been conducted in this paper to achieve the objectives of the project. The detection of the vehicles has been done by using background subtraction. The image produced from background subtraction will be improved by using foreground detection to get better quality of image. Another approach that can be used is by using object segmentation technique which will create a bounding box for each vehicle to increase the accuracy of vehicles detection. On the other hand, vehicles detection can be done by using virtual detection line (VDL) which will be placed perpendicularly to the motion of vehicles. By implementing those methods, the vehicles can be detected and counted automatically, and the classification process of them will be done based on vehicle’s length and width. The results obtained by the system will be compared to the original situation to get higher accuracy rate. The higher accuracy rate will lead to the successful system that can detect, count and classify vehicles automatically.