This work aims at non-destructive detection of surface and subsur-face defects in metallic specimens (e.g. railway track components) via pulsed inductive thermography in a moving setup. A new approach for scanning with changing speed has been developed to enable manual scanning of components in track. A steel speci-men is heated with an inductor, and the surface temperature, including its variation over time, is recorded with an infrared cam-era. The infrared sequence is evaluated by Fourier transform to phase image, in order to increase the image quality. The proposed method uses a referencing object with AprilTags visible in each frame, allowing an adjustment by image registration. Changes in velocity during the scanning process are possible, as demonstrated by a manually performed scan of the specimen and the correspond-ing evaluation. Results obtained with different infrared cameras and scanning speeds were condensed into an estimation of the max-imum scanning speed for a given camera specification.