Pengembangan Aplikasi Klasifikasi Otomatis Plat Nomor Ganjil/Genap

Abstrak

This research aims to develop an application that can classify odd and even license plate automatically. The YOLO and CNN models are trained using a dataset consisting of 1900 images of license plate of belgian cars, and 35000 images of license plate characters. The license plate images has been given annotation using LabelImg. The license plate character images is divided into 35 class, in which each class has 1000 images, these clases consist of the number 0-9 and the letter A-Z with the exception of the letter O. The model is then implemented into an application, for the development of said application, this research uses the incemental model. An accuracy of 99,06% is achieved from the experiment result of the model that is used on this application.