首页> 外文会议>Mexican conference on pattern recognition >A Lightweight Library for Augmented Reality Applications
【24h】

A Lightweight Library for Augmented Reality Applications

机译:用于增强现实应用的轻量级库

获取原文

摘要

To build an Augmented Reality (AR) application it is necessary to recognize a fiducial marker, then to calibrate the camera that is viewing the 3D scene on the marker, and finally to draw a virtual object over the image taken by the camera but in the virtual coordinate system supposed also on the fiducial marker. The camera calibration step give us the transformation matrix from 3D world to 2D on the screen, and the pose of the marker with respect to the virtual coordinate system. An AR application must run interactively with the user, and also in real time. Performing all these calculations in a embedded device such as a Single Board Computer (SBC), a tablet, or a smartphone, is a challenge because a normal numerical analysis library is huge, and it is not designed for such devices. In this article we present a lightweight numerical library, it has been developed thinking in such computing restricted devices. We show results on two AR applications developed for the Raspberry Pi 3 SBC.
机译:要构建增强现实(AR)应用程序,必须先识别基准标记,然后校准在标记上查看3D场景的相机,最后在相机拍摄的图像上绘制虚拟对象,但要在相机上虚拟坐标系也应该在基准标记上。相机校准步骤为我们提供了从3D世界到屏幕上2D的转换矩阵,以及标记相对于虚拟坐标系的姿态。 AR应用程序必须与用户交互并实时运行。在诸如单板计算机(SBC),平板电脑或智能手机之类的嵌入式设备中执行所有这些计算是一个挑战,因为普通的数值分析库非常庞大,并且并非针对此类设备而设计的。在本文中,我们提供了一个轻量级的数字库,它是在这种受计算限制的设备中开发的。我们展示了为Raspberry Pi 3 SBC开发的两个AR应用程序的结果。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号