The Federal Railroad Administration (FRA) has sponsored several programs to help reduce incidents at highway-rail crossings and improve the safety of both vehicles and pedestrians. One such recent project developed an artificial intelligence (AI)-based machine vision system to help inspect highway-rail grade crossings with forward-facing video (FFV) obtained from a Class I geometry car. The goal of this research is to apply the same concepts for locomotive-mounted FFV cameras, currently in common use within the industry; it complements manual inspections. FRA supported this research through its Phase I Small Business Innovation Research (SBIR) initiative.