Analisis Penggunaan Filter Pada Sistem Pengenalan Plat Nomor Menggunakan Phase Only Correlation

Aini, Ossi and Hadi, M. Zen Samsono and Assidiqi, M. Hasbi Analisis Penggunaan Filter Pada Sistem Pengenalan Plat Nomor Menggunakan Phase Only Correlation. EEPIS Final Project.

Preview

PDF - Published Version Download (388Kb) | Preview

Abstract

In parking area of a building or shopping, always found the park entrance are record plate numbers of vehicles on the ticket that function as entry ticket number and the identity of the owners of vehicles that will be examined during the match out of parking area. Currently, digital image processing technology is growing along with the progress of technology. With technology, it is expected of a system of vehicle number plate recorded automatically, so that people just only to capturing the image of plate. The process to record is taking the letters or numbers that are on the plate number, then do image process filter to each character and then used Phase Only Correlation�s method. This method is correlating of two images with the spatial image domain into the frequency domain to obtained a phase�s value. To change the image from spatial domain into the frequency domain algorithm is used 2D-Fourier Transform Discrette. Algorithm is expressed as a spatial image magnitude and phase. Magnitude is expressed how many frequency components in the image. While the phase is expressed where position frequency of the image. After the phase is obtained, then created an array index value is converted into the letters or numbers. After score matching or maximum value of the results is obtained from POC�s method, the matching score which has the highest score will be recognized by the system as the most similar image. The results of this final project is a system can recognize a vehicle number plate with the output is form of text characters from a number or letter plate of vehicles that are detected. Key Word : Image Processing, Phase-Only Correlation (POC), 2D-Discrette Fourier Transform, filtering image.

Actions (login required)