Optimasi Pengenalan Posisi Plat Nomor Kendaraan Menggunakan Kombinasi Metode MSER Dan Harris Corner

The research proposes an innovative approach to identifying and positioning vehicle number plates using a combination of Maximally Stable Extremal Regions (MSER) and Harris Corner methods. The MSER method is used to detect stable regions on the vehicle number plate image. MSER has the ability to recognize areas that have significant contrast intensity, which often represents the characteristic of the number plate. After identifying the potential regions, the Harris Corner method was applied to determine the characteristic angles. The cross points on the number plate. The combination of these two methods allows for more accurate and reliable identification of the position of the number plate. In this study the author performs optimization by changing the preprocessing part and the part of the MSER method. In the preprosessing the author changes the morphological part of a filter, in the section of the method MSER adds input arguments such as ThresholdDelta, RegionAreaRange, and MaxAreaVariation. The results of this study are 99.27% accuracy, 82.73% precision and 83.14% recall. Previous studies were 98.85%, 67.61% and 79.66% recalls. Based on the results of these values, the study has successfully optimized previous studies.