This study presents a real-time refinement procedure for depth data acquired by RGB-D cameras. Data from RGB-Dcameras suffer from undesired artefacts such as edge inaccuracies or holes owing to occlusions or low object remission. In thiswork, the authors use recent depth enhancementfilters intended for time-of-flight cameras, and extend them to structured light-based depth cameras, such as the Kinect camera. Thus, given a depth map and its corresponding two-dimensional image, wecorrect the depth measurements by separately treating its undesired regions. To that end, the authors propose specificconfidence maps to tackle areas in the scene that require a special treatment. Furthermore, in the case offiltering artefacts, theauthors introduce the use of RGB images as guidance images as an alternative to real-time state-of-the-art fusionfilters thatuse greyscale guidance images. The experimental results show that the proposed fusionfilter provides dense depth maps withcorrected erroneous or invalid depth measurements and adjusted depth edges. In addition, the authors propose a mathematicalformulation that enables to use thefilter in real-time applications.