Royale
3D相机SDK
手部追踪

  • Royale SDK
    Royale软件包为飞行时间(ToF)3D相机提供了一个轻量级框架。它既是为pmd 3D相机(如flexx2)量身定做,也让合作伙伴和客户能够在其目标平台上评估或整合3D TOF技术。这减少了首次演示和进入市场所需的时间。
  • Royale包含所有操作ToF的相机所需的逻辑。通过一个高级接口可以方便地控制相机。Royale框架完全采用C++设计,使用C++11标准。

    下载ROYALE

    (需要密码)

隆重介绍
Royale 5 SDK

更精准的数据
更出色的性能

我们最新的主要软件更新Royale 5正式发布,已经可以免费更新。这个新版本包含了一些功能升级,使其获取前所未有的速度和效率。我们还对深度API进行了改动,以改善数据质量和最新3D ToF相机(flexx2等)的处理性能。

API升级

我们鼓励所有用户更新项目代码,以利用全新升级带来的便利。我们专门提供了协助功能来帮助更新。如果您有任何问题或需要帮助,请联系我们的客服团队。对API改动造成的任何不便,我们向您表示歉意。但为了确保能为用户提供最佳体验,这些改动是必要的。

API改动
  • 为我们的3D飞行时间(ToF)相机显著优化处理管道,确保用户获得更好的数据质量。
  • 对三维点云进行更好的过滤和分割,使深度测量更加准确和精确。
  • 为机器人、自动化和安全等领域的应用开辟了新的可能性,使用户能够开发更先进、性能更强大的系统。
  • 改进了API处理能力,获取前所未有的速度和效率。
  • 通过优化,可以在树莓派等计算能力有限的设备上顺利运行。

Royale SDK
4种方式查看深度数据

  • 点云视图
    用彩色x-y-z坐标呈现每个3D像素。
     
  • 深度图视图
    深度数据的彩色呈现。
     
  • 红外线视图
    从红外视觉模式中捕捉的振幅图像。
     
  • 叠加视图
    红外线图像和深度图的结合。

学习更多相关知识

观看我们的教程,学习如何用您的3D相机设置和运行Royale。您还可以了解高级应用,例如ROS2集成或边缘计算以及物体识别。

ROYALE 教程

用户案例
3D推动革新升级

室内空间重建

pmd sensors are a viable solution to locate objects or people inside large environments, such as buildings. This use case is optimized for long range scanning at a maximum data quality. By making use of multiple frequencies the ambiguity range of the sensor signal can be increased by several magnitudes. At the same time this sampling methods leads to an increase in data confidence and applications with very high demands in data quality can be realized.

空间扫描、室内导航

For mapping applications demanding an enhanced situational awareness quick response times are a necessity. These demands are met by increasing the framerate at a minimum cost in data quality.

三维物体重建

Scanning and reproduction of man-sized objects in close proximity demands high data confidence equal to environmental mapping. Since in general the objects of interest are in closer proximity, the range requirements and necessary integration time can be lowered in favor of faster scanning speed.

中等尺寸物体识别、人脸重建

In general the quality demands of applications in the field of pattern and object recognition are less demanding than metrological applications. On the other hand, a quick system response time is mandatory. Therefore the integration time and correspondingly the data quality is lowered in favor of faster framerates.

远程协作、分步指导、桌面游戏

For modern gaming and collaborative applications a quick system response is even more important. Since the range requirement can be lowered and the noise performance of pmd sensors is directly related to the object distance, higher framerates at equal data quality can be realized.

小型物体/产品识别

For hand-size objects and products the necessary range requirements can be further limited and only one scanning frequency is sufficient. Therefore the framerate can be almost doubled and vice versa the overall scanning speed.

手部追踪

The precise detection and recognition of hand gestures in 3D space is very demanding, both in data quality and processing speed. Hence a special use case has been devised offering optimum setting for this special application.

混合模式

The mixed modes can be used to run two or more different modes at the same time, by capturing frames that fit into separate use cases.

For the flexx, Royale offers mixed modes that are a combination of short range/high fps (e.g. for hand tracking) and long range/low fps (environmental scanning). They can be selected like any other use case; to enable the application to distinguish the different use cases the frames will be delivered as separate streams.

操作系统

Royale支持以下操作系统:

  • Windows 10
  • Linux(使用Ubuntu 20.04测试)
  • OS X(使用El Capitan 10.11测试)
  • Android(测试版本为 Android 8)
  • Linux ARM(32位版本使用树莓派Raspbian GNU/Linux 10(Buster)测试,参考日期2020-08-20 64位版本使用Raspberry Pi 4 Ubuntu 20.04 ARM 64测试)

硬件要求

Royale is tested on the following hardware configurations:

  • PC, AMD Ryzen 7 PRO 5850U (64 bit
  • Samsung Galaxy S9
  • Raspberry Pi 4

Also make sure to fullfill the requirements, when working with one of these platforms:

  • On x86 platforms : AVX2 support required
  • On ARMv7A platforms : NEON >= v4 required
记录

软件生态系统