Sidehill and through cuts are often used in the construction of new rail systems to reduce the length, curvature, and grade of the route to increase operation efficiency. Consequently, rights-of-way that utilize cuts are susceptible to damage from slope failure events such as shallow-rapid landslides and rock-falls. At-risk slopes, or geohazards, are traditionally assigned severity grades using slope stability analysis methods derived from intensive field investigations and historical failures. Anticipating slope failures that may occur due to common weather events in the region helps protect railroad assets and ensure safe operations. This research aimed to create a new slope stability analysis method by processing digital images of the railroad right-of-way recorded by inspection vehicles. Computer vision techniques were used to identify and quantify geohazard features that indicated slope instability. Specifically, an object detection algorithm was trained with a set of right-of-way images where features of concern were labeled and located. These features, along with digital elevation maps of the track section, can then be used to develop a relative risk assessment algorithm to grade each geohazard’s landslide likelihood. The resulting algorithm would provide a geohazard risk index that can be determined from publicly available data and video recordings from inspection vehicles and can be used by railroads to mitigate the potential damage from a landslide event.