Noise is reduced in video frames captured by an image sensor (103) of a monitoring camera (100) by using image data representing a captured video frame (302) and image data representing a previously captured video frame (304) at least partially overlapping the captured video frame (302), as well as motion data representing global motion vectors for the captured video frame (302) in relation to the previously captured video frame (304). The captured video frame (302) and the previously captured video frame (304) are spatially aligned using the global motion vectors to provide a set of spatially aligned video frames, and temporal noise reduction is performed on the captured video frame using the spatially aligned set of video frames to provide a temporally noise reduced captured video frame.