Reconstructing objects in three dimensions (3D) has always been one of the primary tasks in the domain of computer graphics (CG). Succeeding in this task in real-time, including texture reconstruction, and with very low-cost hardware is the challenge presented in this paper. Graphical processor unit (GPU) Shader algorithms are the base of our 3D density as well as surface color econstructions. We first compare the different geometrical approaches to cast the object silhouette extracted from video cameras into a large voxel space. To test and optimize material configurations and geometrical models, we then present a novel approach using virtual 3D objects. Finally, we demonstrate the abilities of our model with different comparative tests.