Iraqi Traffic Signs Detection Based On Yolov5
Ammar Abdullah Aggar, Abd Al-Razak T. Rahem, Mohammed Joudah Zaiter · 2021
Traffic signs object detection has gained high interest in recent years, as one of the most significant object detector applications. The development of deep learning technologies gives support to traffic signs detector which it offers several advantages, including the benefit of high detection precision and the timely response to condition changes of traffic signs. Therefore, this paper shows an efficient method for detecting traffic signs. Hence, it implements a new Iraqi Traffic Sign Detection Benchmark (IQTSDB) dataset based on You Only Look Once version 5 (YOLOv5) algorithm. The experimental results show that the implementation of the IQTSDB dataset with YOLOv5 has high efficiency in different conditions such as sunny, cloudy, weak light, and rainy conditions. Besides, real images has been captured for the traffic signs in Baghdad. In addition, the results show that the YOLOv5 has high efficiency in detecting traffic signs of different sizes (small, medium, large), and mean Average Precision (mAP) compared to yolov2, and yolov3.