<![CDATA[Speed ​​limit traffic signs are signs that used to limit vehicle speed. But many drivers break the sign rule because intentionally or unintentionally, resulting in accidents that endanger yourself or others. So, we need a tool that can detect and give a warning to drivers. The system developed in this study detects speed limit traffic signs and perform recognition. The system will detect the color and shape of traffic signs to detect these signs by using color segmentation and Hough Circle Transform. The average accuracy of the system in detecting signs is 94.45%. Next is character detection to separate each character. Character detection is done by color segmentation and Bounding Box for each character. The average accuracy in detecting characters is 88.89%. Furthermore, the character image will be extracted using the division area and the ratio of the foreground pixel to pixel area in each area. The feature extraction results obtained 25 features as input from the Artificial Neural Network to do recognition using the Back-Propagation training algorithm. The average accuracy of the system in doing character recognition is 88.89%. While the average total system computing time is 0.227 seconds.]]>
Copyrights © 2020
