The highway-railroad grade crossings are a hot spot in terms of vehicle-train collisions. The unexpected crossing blockage not only brings traffic congestion but also raises serious safety concerns for commuters. Previous researchers have already investigated the accident loss and frequency of the accidents around the crossing areas. However, there is no dedicated research to facilitate the information exchange between the railroad and street traffic. In this study, researchers at the University of South Carolina initiated the effort of evaluating traffic conditions at the grade crossings and establish two-way communication between the railroad and street traffic to assist the railroad and drivers in their decision making. A customized detection and tracking algorithm based on deep learning was developed. Results presented in this report indicate the traffic during and after the crossing blockage does follow a pattern.