Low-Cost Embedded Solution for Traffic Sign Recognition on Thailand’s Asian Highways

This research aims to develop a low-cost embedded system, specifically a Raspberry Pi 5 board, for speed limit and traffic signal recognition and classification. The dataset consists of images of traffic signs from Asian highways. YOLO model versions 10 and 11, in bothnano and small sizes, were used to classify speed limit and traffic signal signs. The system can detect and classify two types of signs: traffic signal sign and speed limit signs: 80, 90, 100, and 120 kilometers per hour, totaling five signs, and provides an audible alert via a speaker. Experimental results show that YOLO version 11 is more accurate than YOLO version 10 when using the same model size. Additionally, a comparison of speed limit and traffic signal sign detection times across different platforms including a personal computer, Raspberry Pi 4B, and the proposed system demonstrates that the proposed system outperforms the Raspberry Pi 4B in detection speed. This makes it a more efficient alternative f or real-time applications requiring high performance and fast response times.