Traffic Light Control System With Image Segmentation Technique and Pattern Matching Technique Using NI myRIO

Traffic congestion is still a serious problem in many big cities in Indonesia, one of which is due to traffic light settings that are less adaptive to vehicle conditions on the road. This research aims to design and analyze the performance of an adaptive traffic light control system with image segmentation and pattern matching techniques using NI myRIO. The tests include selecting the optimal resolution, comparing fps with webcam input and AVI video, determining the best camera angle, and testing the system in bright, dim, empty streets, and early morning conditions. The results showed that a resolution of 176 × 144 and a camera angle of 90° provided the best performance. AVI video can be used as an alternative in system testing as the results are similar to webcam. The system performed optimally in bright conditions with an F-1 Score of 79.33% and a match rate score of 100% with an average fps of 9.93 fps. However, in dim conditions the performance decreased with an F-1 Score of 69.33% and a Match Rate Score of 41.67% with an average of 10.25 fps. In empty streets and early morning conditions, the system can still operate but sometimes experiences processing delays.