Robust detection method for improving small traffic sign recognition based on spatial pyramid pooling

An extraordinary challenge for real-world applications is traffic sign recognition, which plays a crucial role in driver guidance. Traffic signals are very difficult to detect using an extremely precise, real-time approach practical autonomous driving scenes. This article reviews several object detection methods, including Yolo V3 and Densenet, conjunction with spatial pyramid pooling (SPP). The SPP principle employed boost the Densenet backbone networks extract features. Moreover, we adopt learn features more completely. These models measured compared key measurement parameters such as average accuracy (mAP), working area size, time, billion floating-point number (BFLOPS). Based on experimental results, outperforms state-of-the-art systems. Specifically, obtains 87.8% small (S) target, 98.0% medium (M) 98.6% large target groups BTSD dataset. Our results have shown that highest total BFLOPS (66.111), mAP (99.28%). Consequently, upgrades achievement of all models.