Due to highway real-time monitoring issues, manual monitoring cannot quickly detect abnormalities. Incorrect detection and costly monitoring make it hard to determine traffic's operating state. Experiments show that the YOLOv3 algorithm has improved. Real-time detection improves accuracy. Improving the deep-sort vehicle feature description can speed up vehicle tracking. The system now detects vehicle speeding and congestion in real-time utilizing an upgraded vehicle object recognition, tracking algorithm, and expressway traffic monitoring picture. First, the highway corporation's monitoring system was inspected. After monitoring highway footage in various places, lighting, weather, and other conditions, a comprehensive vehicle data set was developed. A deep learning model for vehicle target detection is built on this data. Second, the labeling tool divides the data set and labels each sample. YOLOv3's border regression loss function increased with the enhanced vehicle detection algorithm. This was achieved through vehicle target identification algorithm research. Build a model of accurate vehicle detection, test the algorithm model, and use the results of the experimental examination of the various loss functions and training data to improve target identification performance and speed. After discussing target tracking, the real-time multi-target monitoring-based deep-sort algorithm will be explained. Trials assess the improved algorithm. If the detection impact is the same, increasing the tracking speed will improve the anomaly detection system's real-time performance. This is true even without boosting tracking speed. Anomaly detection was developed from vehicle target identification and tracking studies. Debugging and analyzing an abnormal event detection system may automatically recognize the speed and congestion of the video vehicle during abnormal occurrences, such as those in expressway surveillance video. It accurately detects real-time anomalous events. Thus, this technology can better track vehicle traffic and predict and identify emergencies. It also supports timeliness and on-site rescue, ensuring lane traffic flows as planned and improving highway business service.