期刊名称：International Journal of Computer Trends and Technology
电子版ISSN：2231-2803
出版年度：2019
卷号：67
期号：6
页码：1-10
DOI：10.14445/22312803/IJCTT-V67I6P101
出版社：Seventh Sense Research Group
摘要：With the development of threedimensional (3D) printing, virtual reality (VR), and augmented reality (AR), a method to accurately determine the structures of 3D objects in various fields including computeraided design (CAD) is required. However, unlike twodimensional structures, analyzing 3D structures is highly problematic in terms of processing speed because the involvement of large number of data. In general, although it is possible to formulate an algorithm based on contact determination to achieve drawing at high speed, it is impossible with such an approach to capture the shape of the structure itself and therefore it cannot be used directly in CAD or 3D printing. Herein, to capture the structure, stereoscopic angles are used to perform internal and external determination for any point in three dimensions. However, the algorithm using the solid angle can be accurately internal and external determination, whereas a method using a conventional central processing unit with a simple triangular function is very problematic in terms of speed. From the above, to focus on the effectiveness of the parallel arithmetic when performing the internal and external determination in accordance with the algorithm using the stereoscopic angle, this study proposes a highspeed determination method using generalpurpose graphics processing unit (GPGPU). In addition, by extending the idea of the 3D angle, it is shown that the shape can be captured accurately and quickly even for a complex shape (a 3D object including layering) inside another complex shape.
关键词：CUDA; Parallel Computing; GPGPU; Internal and External Analysis.
