With the rapid development of UAV technology and the application of 5G communication technology, smart transportation systems are becoming an important tool for solving problems such as traffic congestion and accident rescue. This study designs a UAV smart traffic rescue system that integrates multi-source perception and 5G communication, aiming to improve the response speed, execution efficiency and accuracy of traffic rescue tasks. The system integrates multiple sensors such as vision, infrared, and radar to achieve real-time perception and data collection of the accident scene, and uses the 5G network to transmit data to the command center in real time to ensure rapid decision-making and response. The experimental results show that in conventional rescue systems, the task completion time is reduced by an average of about 30%, and the response speed is increased by more than 40%. Through the fusion of multi-source perception data, the recognition accuracy of the system in complex traffic environments is improved by 15%. The system can perform real-time data transmission in extreme environments, and greatly reduces transmission delays with the support of 5G networks, ultimately achieving efficient execution of accident rescue. Combining multi-source perception and 5G communication technology, the system proposed in this study provides a powerful technical framework for the future application of UAVs in the field of smart transportation, with significant innovation and broad application prospects.