Moving Object Detection and Counting in Traffic with Gaussian Mixture Models and Vehicle License Plate Recognition with Prewitt Method

Detection and counting of moving objects and recognition of license plates are important processes in traffic surveillance. In this study, the development of moving object detection and counting and license plate recognition algorithms are explained. Gaussian mixture models are used to detect, track and count the moving objects in a video sequence. The algorithm shows the total number of moving objects on the left corner of the processed video frame. Prewitt operator and optical character recognition are used to recognize vehicle's license plate. License plate recognition algorithm can recognize license plates of different types of vehicles in different positions without any character limit. Moving object detection and counting algorithm is tested using different videos of moving objects. Then, license plate recognition algorithm is tested using various photos of the different types of vehicles. It could be evaluated that moving object detection and counting algorithm easily detects and counts the moving objects and vehicle license plate recognition algorithm clearly recognizes license plates of the cars.