Super-Reality Systems based on 3D-Computer Graphics

近年来，我们有很多机会在各个领域使用3D数据，并一直期待有助于轻松理解3D数据的技术。最有效的方法之一是将三维数据可视化为三维计算机图形。三维可视化极大地帮助人们从海量数据中理解准确的信息。其中一个非常需要这些技术的领域是环境评估，如风景投影系统。这些字段对于使用真实场景和虚拟对象合成3d图像的方法是必要的。许多研究者一直在研究3d图像合成系统。典型的技术分为四种方案；阴影，阴影，分形和纹理映射。将真实景物与计算机图形学相结合的图像合成系统是其中的典型技术之一。在真实景物与计算机图形相结合的三维图像系统中，我们提出了一种测量真实物体的位置和形状以及获取真实物体的CAD数据的新策略。该策略为人类合成三维图像的CAD数据释放了大量的工作。此外，它还可以节省3d图像渲染所需的大量计算时间。通过标记实景图像的各个域并测量每个域的中心位置来估计实景物体的位置。通过获取域的中心位置，得到域与相机之间的距离。根据到相机的距离，对真实物体的所有域进行排序。虚拟对象与真实对象相似的划分域组成。这些域中离相机最远的域被绘制在屏幕上，以此类推。取消的呈现过程通过在屏幕上绘制域来执行。为了利用三维计算机图形学实现超真实感，我们提出了一种从真实图像中获取纹理并在三维cad数据上自动映射纹理的新方法。可以看出，自动纹理切割可以节省大量的工作，自动纹理映射实现了3d计算机图形的超真实感。然而，这些策略只适用于形状简单的3d物体。将该策略应用于复杂形状的物体是未来的工作方向。少

In recent years, we hava many opportunity to use 3D data in various fields and have been expecting the technologies that help understand 3D data easily. One of the most efficient method is visualize 3D data into 3D computer graphics. The 3D visualization much helps human beings to understand exact informations from a huge data.One of the fields to extremely required these technologies is the environment assessment to such as scenery projection systems. These fields are necessary for a method to composite 3D-images using real scenery and virtual objects. Many researchers have been studing 3D-image synthesizing systems. Typical techniques are classified into four schemes ; shading, shadowing, fractal and texture mapping.As one of the typical techniques, an image synthesis system using real scenery and computer graphics has been expected. In the 3D-image system using both real scenery and computer graphics, we have proposed a new strategy to measure positions and shapes of real objects an … More d to obtain CAD data of real objects. The startegy releases human beings much works to make CAD data for synthesizing 3D-images. Further, it can save a lot of computation times required for rendering of 3D-images.Positions of real object is estimated by labeling each domain of real image and measuring the center position of each domain. A distance between a domain and the camera is derived by obtaining a center position of a domain All domains of real objected are sorted according to the distance from the camera. Virtual objects to be composed with domains which real objects are divided into similarly. The most far domain from the camera among those domains is painted over the screen and so on.The canceled rendering process is executed by painting domain over the screen.To develope the super-reality using 3D-computer graphics, we have proposed a new method to obtain textures from real images and to map the textures on 3D-CAD data automatically. It is seen that the automatic texture cuting can save much works and the automatic texture mapping achieves super-reality for 3D-computer graphics.Howevere, these strategies are only applied to the 3D-objects with simple shape. It is a future work to develop the strategy applying to objects with complex shapes. Less