A kinect-based application for modelling three-dimensional objects

It is very difficult to generate solutions to reconstruction the three-dimensional real world objects in the form human eye sees by reorganizing their content with the aid of two-dimensional image structure. There no necessity for the human eye to perform a conversion from 2-dimension (2D) to 3-dimension (3D) in respect of its visual quality; this characteristic is an extraordinary ability for the human to recognize the visual world. In order to understand the logical approach of the visual world, such abilities of human are imitated for computer vision, virtual reality, object recognition, augmented reality, restructuring. However, it is fairly troublesome to reconstruction the real world objects in the computer medium in the form of 3D. Basically, realization of a new interaction between the physical world objects and virtual space by utilizing various methods lies behind the background of this thesis study. From this viewpoint, purpose is to provide visualization of 3D model of the real world object in virtual spaces to be constituted by means of a standard 3D camera that has the capacity to generate a scanning of 360 degrees animatedly in order to reconstruction a closed stage in detail. Studies started by making investigation about the review of existing methods on multiple entire scanning in the stage of obtaining overall depth inputs of the object in order to shape the physical objects in a solid structure of 360 degrees. In order to obtain the best results during restructuring of world objects with point cloud, we focused on processing of these points in this study by utilizing Microsoft Kinect and open-source code point cloud library. Consequently, various filtering methods to be applied to the object and its backgrounds were reviewed from real world resident objects on planar surfaces over to depth scans of point cloud library (pcl) taken more than once in animated form; they are visualized in MeshLab and OpenGL visualization platforms after these inputs are combined by triangulation method.