In urban area, frequent train services during peak hours often keep interrupting roadway traffic approaching level crossings (LXs). In Taiwan, more than 24 trains per hour pass the LXs in peak hours, and it leads to more than 40% time blocked by the barriers. To mitigate the potential risk of accidents between trains and roadway vehicles, almost 20% LXs had been installed with obstacle detection devices. It, however, leads to the following issues: (1) when the level crossings open and then close within a short period of time, the vehicles would likely be trapped in the dangerous area; (2) if the closing time at level crossings is not adjusted effectively, it would seriously impact the traffic flow and increase the possibility of trespassing; (3) the false alarm triggered by the obstacle detection devices has serious impact on train punctuality. To mitigate the influence on roadside traffic and improve the safety, this study developed an intelligent module which consists of the following functions: (1) transmitting real-time videos of level crossings to the approaching trains, (2) constant alarm time for trains with different approaching speeds, and (3) extending alarm time for trains from the opposite directions within short time periods. A level crossing located on curved line in urban area was selected to test the performance of these functions in an actual scenario. Alternative solutions were also adopted in the tests to compare the performance for reference. The results show that the module can be viable in practice, but needs further works to ensure the safety, such as developing a fail-safe mechanism and overcoming the software security issues.