The Application of Template Matching Algorithm to Improve Accuracy of Plate Number Recognition

Abstract

Pattern Recognition is a part research of Image Processing. In term implementation of Pattern Recognition
used for variety research needs such as Identify Transportation License Plate Number. Identifying license
plate number is one process to reduce crime, accident, and support related agencies to process information
traffic data. Taking picture of license plate number by camera create image that can be processed by Optical
Character Recognition (OCR). OCR is conversion process of picture that contain alphabet and numeric
object into ASCII character. In creating of OCR, there’s many kind of algorithm, we used Template
Matching algorithm as main research. Process of Template Matching algorithm is fitting all part of picture
taken by camera with image that has been processed into Template Data. However like another algorithm,
this algorithm had a deficiency about accuracy and time to recognition process. To create and enhance of
accuracy in Template Matching algorithm, we create 3 scenarios to take a picture for compare the result of
accuracy and time to OCR process also comparing Template Matching algorithm with Learning Vector
Quantization (LVQ) as accuracy comparison method. Based on result of Template Matching algorithm had
84% accuracy and 0,942 Milliseconds to process OCR.