Although deep learning-based object detection has produced excellent performance, there are still many issues with images from real-world capture, including rotating jitter, blurring, and noise deletion. The impact of these issues on object detection is significant. The main goal of this paper is to develop a real-time “You Only Look Once” (YOLOv3) algorithm to detect traffic signs. Compared to all other object detection algorithms, the YOLO method has a number of benefits. In contrast to different algorithms, YOLO looks at the image entirely by making predictions of the bounding boxes utilizing a convolutional neural network (CNN), determining the probability of each class for these boxes, and detecting the image more quickly. The proposed method applies a single neural network to the entire image. Then this network divides that image into regions, which provide the bounding boxes and also predict probabilities for each region. These generated bounding boxes are weighted by the predicted probabilities. The proposed method achieves 99% accuracy in the detection process.