| The current project presents a method based on image processing techniques for the identification of moving vehicles as they approach a signaled intersection. A set of fixed cameras located before the intersection monitor the street continuously, taking pictures of the approaching object. A feature extraction algorithm is presented, which identifies a set of features in the image, and calculates a distance metric, measuring the difference of the current image from images stored in a database of vehicles. If the calculated distance metric is very small, then the present vehicle is successfully classified as being the same type as one of the vehicles stored in the database. A successful application of this method was implemented using real-time data. In the particular application presented in this project the vehicle to be identified is an ambulance car, whose images have been previously stored in the vehicle database. |
