《计算机应用》唯一官方网站 ›› 0, Vol. ›› Issue (): 192-200.DOI: 10.11772/j.issn.1001-9081.2023081097

• 多媒体计算与计算机仿真 • 上一篇    下一篇

基于ARM平台的VR头部追踪系统的设计与实现

张向军, 黄国书, 邱涛(), 王明, 尹逊刚, 范兵兵   

  1. 歌尔科技有限公司,山东 青岛 266000
  • 收稿日期:2023-08-14 修回日期:2024-03-21 接受日期:2024-03-25 发布日期:2025-01-24 出版日期:2024-12-31
  • 通讯作者: 邱涛
  • 作者简介:张向军(1981—),男,河南安阳人,高级工程师,硕士,CCF会员,主要研究方向:虚拟现实、系统软件
    黄国书(1993—),男,台湾台中人,工程师,硕士,主要研究方向:机器视觉、SLAM算法
    邱涛(1990—), 男,山东青岛人,工程师,硕士,主要研究方向:虚拟现实、系统软件
    王明(1987—),男,山东青岛人,高级工程师,硕士,主要研究方向:虚拟现实、系统软件
    尹逊刚(1990—), 男,山东济南人,工程师,硕士,主要研究方向:三维数据可视化
    范兵兵(1998—),男,山西太原人,硕士研究生,主要研究方向:SLAM算法。

Design and implementation of VR head tracking system based on ARM platform

Xiangjun ZHANG, Guoshu HUANG, Tao QIU(), Ming WANG, Xungang YIN, Bingbing FAN   

  1. Goertek Company Limited,Qingdao Shandong 266000,China
  • Received:2023-08-14 Revised:2024-03-21 Accepted:2024-03-25 Online:2025-01-24 Published:2024-12-31
  • Contact: Tao QIU

摘要:

随着虚拟现实(VR)一体机的蓬勃发展,内向外六自由度(6DOF)的追踪技术已成为主流的头部追踪方案。然而,由于要同时处理大量视觉及惯性测量单元(IMU)数据,计算量较大,在ARM移动平台计算能力有限的情况下,如何同时达到较好的追踪精度和稳定性成为实现该系统的关键技术难点。为此,基于ARM移动平台和开源ORB-SLAM3算法,设计并实现一套追踪精度高、稳定性好的VR 6DOF头部追踪系统。首先,通过高精度机械臂对IMU和Camera进行联合标定,并通过IMU积分信息与VIO(Visual-Inertial Odometry)进行位姿融合优化;其次,结合VR使用场景,调低单帧图像的特征点规模,以降低算法复杂度,达到近1 kHz的6DOF位姿输出频率;最后,提出一种针对视觉跟踪丢失的快速处理策略,以提高追踪状态的稳定性。经过以上优化后的算法运行在高通XR2平台上的实验结果显示,相较于传统的ORB-SLAM3算法,所提算法在相对位置误差(RPE)和相对旋转误差(RRE)指标上分别减小了44.4%和73.4%,在实时性和精度指标上能满足VR一体机需求。

关键词: 虚拟现实, ARM平台, 六自由度, ORB-SLAM3, 惯性测量单元

Abstract:

With the rapid development of Virtual Reality (VR) all-in-one devices, inside-out Six Degrees of Freedom (6DOF) tracking technology has become the mainstream head tracking solution. However, these tracking algorithms need to process large numbers of images and Inertial Measurement Unit (IMU) data simultaneously, which consume a lot of computational resource. When deploying the algorithms on ARM mobile platform, achieving good tracking accuracy and stability with limited computing power has become a key technical difficulty. Therefore, based on ARM mobile platform and ORB-SLAM3 algorithm, a VR 6DOF head tracking system with high tracking accuracy and stability was designed and implemented. Firstly, the joint calibration of IMU and Camera was performed by the high-precision manipulator, and pose fusion was optimized through IMU integration information and VIO (Visual Inertial Odometry). Then, to adapt to the VR usage scenario, the scale of feature points in a single frame image was decreased to reduce the complexity of the algorithm, achieving nearly 1 kHz the 6DOF pose output frequency. Finally, a fast-response strategy for visual tracking loss was proposed to improve the stability of the tracking state. After the above optimizations, experimental results of the proposed algorithm on Qualcomm XR2 platform show that compared with the traditional ORB-SLAM3 algorithm, the proposed algorithm has the Relative Position Error (RPE) and Relative Rotation Error (RRE) indicators decreased by 44.4% and 73.4% respectively. In terms of real-time performance and precision, the proposed algorithm satisfies the needs of VR all-in-one devices, as demonstrated by the experimental results.

Key words: Virtual Reality (VR), Advanced RISC Machine (ARM) platform, Six Degrees of Freedom (6DOF), Oriented Fast and Rotated Brief-Simultaneous Localization And Mapping Three (ORB-SLAM3), Inertial Measurement Unit (IMU)

中图分类号: