3D ROOM Reconstruction
Abstract
3D objects or elements reconstruction has become a popular problem in the last years. It became relevant in the industrial world as in peoples' everyday lives. Among the pioneering examples, one can notice 3D printing, augmented reality games and simulations and even medical imaging.
There are various solutions for the 3D reconstruction problem, based on different assumptions made about world structure’s attributes. Each one of them has its challenges and obstacles and its own pro's and con's. There is no right or wrong solution.
Among other things, the different methods offering 3D reconstruction solution differ one from each other by the raw data they are analyzing. Is the raw data depth imaging, have the pictures been taken from one point of view or several, etc.
This project handles the problem of reconstructing an inner space, a room, focusing on maximizing the automation of the process, considering the images had been taken from several points of view in the room.
The solution of this project includes improving the given images for the next step, generating a point cloud from those images and performing geometrical analysis in order to produce the final output. Those steps will be discussed later in this summary.