Deteksi Objek Plat Nomor Kendaraan dengan Metode CNN

Abstract

The public's need for transportation to date is very high, this can be seen in front of the number of vehicles, both private and public vehicles that go back and forth from rural to urban areas. This results in congestion due to the density of vehicles as well as less organized parking management. On the other hand, with increasing population growth, land becomes narrower, while public interest in buying vehicles, both two-wheeled and four-wheeled, is increasingly inevitable, as a result of the increasingly affordable prices of motorized vehicles. According to research (Wini Mustikarani & Suherdiyanto. 2016:1), one of the factors that cause congestion is indiscriminate parking activities. There is also parking management that is currently being carried out using manual methods, such as writing or typing manually to record motorized vehicle numbers. To minimize manual work, one innovative way is to apply artificial intelligence as image processing with deep learning that utilizes an artificial neural network using the Convolutional Neural Network (CNN) method. By conducting training and testing on images of Indonesian license plates of motorized vehicles, machine learning will do its job like humans who can recognize the object of the number plate of a motorized while and record the vehicle number for further analysis, both for parking data purposes and data from the Department of Transportation, and the Police.