Design of Roadside Vision Perception Algorithm for V2X Technology

With the continuous development of autonomous driving technology, many experts and scholars pay more and more attention to V2X technology. Roadside perception technology is particularly important for autonomous driving perception fusion system based on V2X technology. Roadside vision perception technology is the core of roadside perception technology. Based on the improved YOLOv5 targets detection algorithm and DeepSORT tracking algorithm, accurate and fast perception of roadside targets is realized. It effectively improves the accuracy of targets detection and provides a large number of useful information for autonomous driving perception fusion system. It ensures the safe driving of V2X autonomous vehicles, which is of great significance to the realization of V2X high-level autonomous driving technology.