Speed limit traffic sign detection and recognition based on support vector machines

Abstract

Speed limit traffic sign recognition plays a key role in intelligent transport system (ITS), especially in driver assistant system (DAS) and intelligent autonomous vehicles (IAV). Although traffic signs are clearly defined in color, shapes for easily detecting purpose, an excellent traffic sign detection system still be a challenge for researchers and manufactures because of the strict requirements of correct rate to be able to apply in the reality. In this paper, a speed limit traffic sign detection and recognition program was developed based on Visual Studio and OpenCV library. We used color probability model to detect the candidates and applied Histogram of Gradient (HOG) combining with Support Vector Machine (SVM) classifier to remove all wrong candidates, keep only speed limit traffic signs. Then the information of speed limit traffic sign was extracted. The testing results show that the system can detect and recognize the information of limit speed traffic signs with high correct rate even in complicated background conditions and existing overlapped area in traffic signs.