Embedded System for Real-time Traffic Signs Recognition with Augmented Reality Implemented in ARM

Automatic detection and recognition of traffic signs improve driving safety by alerting drivers to hazards they may not perceive. This paper proposes an embedded real-time methodology for traffic sign detection and recognition, enhanced with augmented reality to increase driver vigilance. The algorithm operates in three stages: detection, recognition, and augmented reality merging. Color filtering and shape matching are used to extract Regions Of Interest (ROIs), followed by a Haar cascade classifier for detection and categorization. The recognition stage utilizes the ORB algorithm. Implemented on an ARM platform, the system achieves 96 % accuracy in detection and recognition at 12 fps in real-world conditions.