Aryo Ronaldo Hamba Pulu Pengarang YULIANTO TRIWAHYUADI POLLY Dosen Pembimbing 2 ADRIANA FANGGIDAE Dosen Pembimbing 1 Yulianto Triwahyuadi Polly Penguji 1 Adriana Fanggidae Ketua Penguji Derwin Rony Sina Penguji 2

One type of traffic sign is a warning sign that is used to warn road users. At this time there are still many road users who do not pay attention to the presence of warning signs, which can harm and endanger road users, so it is necessary to have a system that can detect the presence of warning signs on the road. Warning signs can be detected based on color and number of angles because in general warning signs are yellow and are rectangular and rhombic. The detection process is carried out using the Color Filtering method to filter colors, the BLOB Counter method to search for interconnected pixels, and the Simple shape checker to check the shape of the object. At the color filter stage with Color Filtering, 2 color parameters are used, namely the dark color parameter with a value of R(255,75), G(255,75), B(75,0) and the light color parameter with a value of R(255,85), G(255, 85), B(80,0). The system was tested on 120 primary data images and 30 secondary data images. Primary data collection was carried out in the afternoon and evening, with a distance of 5-10 meters and 10-15 meters. The average accuracy given by the system in testing on primary data is 88,5%, and secondary data is 90%, so the overall average accuracy is 89.2%. The conditions of time and distance of taking that provide the highest accuracy occur in daylight conditions with a distance of 5-10 meters. System failure in detecting the state of warning signs on the road is due to insufficient or excessive lighting conditions at the time of image capture, the angle of shooting is too tilted, and adjacent to objects that have a yellow color.