In this paper we propose a methodology to identify and label the structure components of a typical indoor environment, in order to generate a semantic model of the scene. The structures that are able to be identified are: walls, ceilings, floors, doorways with open doors, doorways whose closed doors are not at the same level of the wall and windows partially occluded. These elements should be flat, in case of walls, floors and ceilings, and with a rectangular shape in the case of windows and doorways. The identification of these structures is figured out through the analysis of the contextual relationships among them. The relationships taken into account are parallelism, orthogonality and structure¿s position in the scene. Point clouds were acquired using a RGB-D device, the Microsoft Kinect Sensor. The obtained results were highly satisfactory; a precision of 99,03% and a recall of 95,68% was achieved, which shows a great contribution to the solution of this kind of challenges.