Identifikasi Karakter Pada Plat Nomor Kendaraan Menggunakan Ekstraksi Fitur ICZ dan ZCZ dengan Metode Klasifikasi K-NN

Abstract

License plate is a unique code consisting of letters and numbers arrangement. Each vehicle has a license plate that serve as the identity code of the vehicle itself. Since license plates are unique, it often used for recording, such as in the parking area. The character identification characters on the license plates are made automatically using the input in the form of license plate images to facilitate the recording process. K-NN classification and feature extraction ICZ and ZCZ used in the identification of character identification. ICZ and ZCZ feature extraction obtain features by computing the character centroids and divide the image into n equal zones then computing the average distance from the character centroid to every pixel in the image. The character identification accuracy using 14 zones and ICZ are 97.00% while using ICZ+ZCZ feature extraction are 98.17%. The license plate identification accuracy using feature extraction plates ICZ+ZCZ are 72.31% for zero tolerance and 89.23% for one fault tolerance.