Development of a Real-Time Object Detection System in the Highway Environment Using the YOLOv5 Algorithm for the Blind

Individuals with visual impairments face significant challenges in navigating daily life, particularly in identifying objects on the road, such as vehicles, which affects their safety and mobility. Previous research has proposed various solutions using object detection technologies, yet there remains a need for further development, particularly in enhancing real-time performance, accuracy and practical deployment. This study aims to develop a real-time object detection system in the highway environment using the YOLOv5 algorithm for the blind deployed on web, specifically designed to support visually impaired individuals. Testing involved analyzing the latency and frames per second (FPS) of the detection system, crucial for real-time performance. The results indicate an average total latency of 1.711 seconds, a detection latency of 0.303 seconds, and an FPS of 1.65, with a Mean Average Precision (mAP) of 0.933 at an IoU threshold of 0.5, underscoring the system's capability to operate efficiently in real-time scenarios. The system achieves high accuracy and low latency real-time performance, demonstrating the effectiveness of YOLOv5 in practical applications.