In this project, a mobile phone application will be developed which is aimed at creating a 3D model of a room based on the RGB images taken in a sequence by the user. As the first step, the user will move around the room, and capture the images at a prescribed rate. Throughout the latter process, the user will be instructed by means of warnings preventing them from moving in wrong directions or with problematic orientations. Next, the images will be automatically uploaded into an online cloud, where the server will process them in order to create the desired 3D model. The foregoing procedure will consist, first of all, in finding out the estimated camera poses using stereovision. The underlying principle leading to the most efficient implementation of the geometric calculations will be determined based on the results of prospectively investigating the literature intending to decide on the fastest and most efficient way of doing so. Subsequently, the corresponding depth maps will be created using the information obtained from the aforementioned process, which will be preprocessed and denoised, and then merged, in order to reconstruct an initial point cloud comprising all the frames included in the sequence. Finally, post processing and finalizing the model will be applied, which should involve correcting the usual global consistency error, smoothing the outcome and triangulating the refined 3D point cloud, leading to the ultimate 3D model.