Evaluation of object detectors in recognizing crossroad intersection triangle sign

Abstract

Variations in perspective, illumination, occlusion, motion blur, and weatherworn degeneration of signs could all be crucial in identifying road signs. The goal of this work is to evaluate cascade object detector and Speed Up Robust Features in detecting, recognizing crossroad intersection triangle sign, and determining the optimum threshold value. The current work is executed in Baghdad's streets during the daytime. Results showed the effectiveness of cascade detector in detecting the triangle sign than SURF with precision lies in the range (0.9-0.98). Finally, the highest precision was recorded at fifteen and twenty-five threshold values for cascade and SURF approaches respectively.