A scene is monitored by receiving image data representing a thermal image captured by a thermal image sensor (13) and a visual image captured by a visual image sensor (15), the thermal image and the visual image depicting a same view of the scene. Objects (9) are detected in the visual image and in the thermal image. A mask (17a) is defined in the view of the scene, and it is determined if an object (9) or an object part, inside the mask (17a), is present in both the thermal image and the visual image. If so, a modified mask (17b) is provided by excluding an area occupied by the object (9) or the object part from the mask (17a).