Statistical Versus Transform Methods: An Experimental Comparison of Vehicle Plate Location Detection for a Multi-Camera Tracking System

The aim of this study is to construct a multi-camera tracking system for vehicle plate recognition. Recognizing a car plate on a highway is more challenging task than recognizing a vehicle plate at a toll-gate or a parking. This is because, at the parking and the toll-gate, the camera has already fixed to capture an image of a car plate. But, on the highway there are many cameras installed to capture images and every camera has different angles of images. As a result, the images are captured in varied imaging conditions and not focusing on the vehicle plate itself. Beside that, the cars on the highway are speedy, and sometimes the weather is not good. Therefore, the captured images are blur and not clearer. Based on those situations, we need a system that is able to recognize the location of a vehicle plate first. In this paper we propose two different methods for vehicle plate location detection. First, we used segmentation method and second, we applied fast Fourier transform and histogram method to recognize the location of a vehicle plate. The result was discussed and the methods are compared at the end of this paper.