Correlation methods of OCR algorithm for traffic sign detection implementable in microcontrollers

Abstract

This paper focuses on the correlation methods applicable for the recognition system of speed limit traffic signs and correlation methods comparison. The correlation method is one possible manner of the OCR algorithm (Optical Character Recognition) used to determine the degree of similarity between the input matrix and the defined pattern matrix. The presented correlation methods are verified using the proposed comparison algorithm, where the output data are evaluated by statistical methods of the exploratory statistic data analysis. The part of the recognition system for OCR algorithm proceeds is very time consuming and the limitation of microcontroller type depends on frequency instruction processing. High accuracy of the recognition system can be achieved by increasing the resolution of camera system, by segmentation methods of input image signal, correlation method type. © 2013 Springer-Verlag Berlin Heidelberg.