For a light rail system, smooth contact between the vehicle and the guide beam is critical for reducing the friction and the vibration of an operating vehicle. Therefore, the shape of guide beams needs to be controlled with mm-level accuracy during the construction. Currently, most methods for detecting shape defects of guide beams, such as experimental run of vehicles, are costly and tedious, and can only identify defects after the completion of the construction, causing reworks and delays. From dense point clouds collected by laser scanners, inspectors can manually extract geometric features and conduct virtual inspections of guide beams. However, the manual geometric feature extraction process impedes effective utilization of point clouds for the shape analysis of guide beams. This research developed a semi-automatic approach for simultaneously extracting the axis parameters (e.g., radius) and cross-section features (e.g., width) of a guide beam using a Hough-Transform based approach, and discusses factors (e.g., data density) influencing the performance of this approach.