The invention relates to a method of processing a video, including: receiving a video of a scene being monitored; generating a second video, including one or more video frames, which non-photorealistically represents the scene being monitored; analysing at least the video of the scene being monitored to determine the presence of a threat within the scene; and in the event that a threat is detected within the scene, determining metadata in relation to the threat and representing said metadata in the second video.