Jurnal Infra
Vol 7, No 1 (2019)

Pengenalan Karakter pada Plat Nomor Indonesia dengan Tilt Correction dan Metode Faster R-CNN

Kevin Nyoto Susanto (Program Studi Teknik Informatika)
Kartika Gunadi (Program Studi Teknik Informatika)
Endang Setyati (Program Studi Teknik Informatika)



Article Info

Publish Date
15 Jul 2019

Abstract

The growth of the number of vehicles in Indonesia is very rapid and has caused many problems, such as the problem of parking security systems and access control of vehicles. With the development of technology, vehicles’ license plates can be detected automatically by a system with the help of Digital Image Processing and Artificial Neural Network.This study uses the Canny Edge Detection method to detect license plate objects in the image. Before classifying characters on license plates, the perspective distortion in the image can be removed using the Planar Homography method. Faster R-CNN is used to detect the position of the car in the image and classify the characters on the license plateThe results of this research program will detect the character of license plates and the color of car in the image. From the test results using the researcher’s dataset, the accuracy of detection of the characters in the license plate reached 82.14% and the accuracy of detection of the cars’ colors reached 78.54%.

Copyrights © 2019