Towards Automatic 3D Reconstruction from 2D Floorplan Image

Reconstruction of 3D model representation from a 2D image(s) is proven to be a di cult task. In this paper, we present a simple solution to the restricted case of generating model description of a building solely from its 2D oorplan. The motivation of this research is that i) a lot of existing buildings has only the oorplans remained, ii) In general a lot of e orts i.e human labor work, have to be made in order to build the 3D model. We proposed two approaches with di erent degree of automation: The rst approach operates by an user analyzes and breaks down the oorplan into a number of building primitives. The coordinate information are then manually recorded and input to a program which will generates the 3D model descriptions i.e. a popular 3D model le such as 3D Studio etc. Another approach applies image processing techniques in which the user selects and changes the lines in the oorplan into di erent color coded areas. These areas again represent di erent 3D primitives similar to the rst approach. A raster to vector program then converts these primitives into 3D format. The latter approach involves only minimal human e ort. Using our approaches, an ordinary user can produce a 3D model of a building in a matter of seconds. We present some of the results and discuss the relative merits of the two approaches. The presented algorithms are well suited for data generation in architectural walkthrough on old buildings and 3D games development. keyword : 3D modeling, block-based approach, raster to vector conversion, color-coded blocks