Vehicle Number Plate Identification Using Template Matching Algorithm for Automatic Parking System

Nowadays, some parking system in Indonesia still use manual system that is the parking officer manually records every vehicle number plate that will be parked. This process is less efficient, because it consumes a lot of time and prone to errors. The application of digital image processing methods to automatic parking systems can overcome these problems. This research builds an automatic parking system by applying template matching algorithms. Template matching algorithm is used to help the process of analyzing all forms of character image objects on vehicle number plates, which include mapping the pixel intensity of character images, calculating errors, and searching for minimum error values. The advantage of template matching algorithm is that it processes data in the form of matrices whose computation level is not complex so it does not require a long processing time. Thus, the template matching algorithm is expected to be in accordance with the characteristics of an automatic parking system that will process large amounts of data. System testing has been carried out using 160 datasets of vehicle number plates and obtained good results with the highest accuracy of 91.7% and the average processing time of 13.7 seconds.