Home > Published Issues > 2020 > Volume 8, No. 2, June 2020 >

Combining Multiple Feature for Robust Traffic Sign Detection

Fitri Utaminingrum, Renaldi P. Prasetya, and Rizdania
Computer Vision Research Group, Faculty of Computer Science Brawijaya University, Indonesia

Abstract—Traffic sign detection and recognition as one of the digital image processing areas has been conducted in many researches for improving the safety of driver and even make driver more comfort. Many driver are inattentive and underestimate the traffic signs on the road that affect their safety. So that this system can serve as a warning in driving on the highway. In this paper, we apply our proposed method for the detection and recognition of traffic sign. In the detection process is performed by using a method based on color, considering signs have differences with the other objects in terms of color. While in phase of recognition, our approach consist of 3 main schema. To strengthen the recognition process, we implement corner detection using Harris method (HCD), we also implement edge detection using Canny method, and comparing pixel ratio for each traffic sign. All of these schema series were used for feature extraction and the feature data matching using K Nearest Neighbor (KNN). From the experiment by using our proposed method improve the accuracy significantly reach to 90.85 % and also speed up the computational times.

Index Terms—traffic sign, Harris corner, edge detection, pixel ratio, KNN

Cite: Fitri Utaminingrum, Renaldi P. Prasetya, and Rizdania, "Combining Multiple Feature for Robust Traffic Sign Detection," Journal of Image and Graphics, Vol. 8, No. 2, pp. 53-58, June 2020. doi: 10.18178/joig.8.2.53-58

Copyright © 2020 by the authors. This is an open access article distributed under the Creative Commons Attribution License (CC BY-NC-ND 4.0), which permits use, distribution and reproduction in any medium, provided that the article is properly cited, the use is non-commercial and no modifications or adaptations are made.