Traffic Sign Recognition using Blob Analysis and Template Matching

Abstract

A simple, easy to implement algorithm for traffic sign detection based on thresholding, blob detection and template matching has been discussed in this paper. A small portion of theoretical background is also presented about each of the processes used. The result of each process is shown step by step for a sample test image and is discussed in the results section along with the final results. The further enhancements that can be implemented are also presented. The results obtained showed an accuracy greater than 88% when tested with a pool of images specifically containing extreme cases such as low resolution images, skewed images, etc.