DETEKSI NOMOR KENDARAAN DENGAN METODE CONNECTED COMPONENT DAN SVM

Abstract

Research on License plate detection system will be developed in this research. Vehicle plates detection will use a camera that mounted on parking gate. Detection process starts with collecting data using the camera. Preprocessing includes rescale the image into (640.480) pixels, converting the image into grayscale. Smooting process with blur method is used to reduce the image noise. Edge detection using Sobel on the vertical direction and thresholding. Segmentation process will combine 2 methods, Morphology Filter and Connected Component. Support Vector Machine (SVM) is used to test whether the candidate is a plat plate or not. Parking System Testing obtained the following results: Parking System Accuracy in detecting motorcycle plate is 78% and the accuracy car plate detection is 78%

Key words— nomor plat kendaraan, Filter Morfologi, Connected Component, Support Vector Machine (SVM).