The transition from flat, two-dimensional representations to models in three dimensions takes a lot of computing power. For computer aided design this transition was made long ago. Today there is hardly a machine or building designed with two-dimensional drawings. The requirements for computing power will be even higher when augmented reality (AR) applications are used to represent the combination of moving virtual objects and the real environment in real time. Today current smartphones have processors that are far superior to the powerful PCs of ten years ago. But even so, a sophisticated strategy is needed to make AR applications run on the smartphone and to ensure a smooth display.
Smartphones using Android as an operating system that run AR applications are normally based on the platform Tango Google has developed in cooperation with different partners. Besides many other sensors, the time-of-flight based 3D camera from pmdtechnologies is one of the core elements of the platform. The sensors, in particular the 3D camera, generate a very large data volume. A three-dimensional point cloud represents the environment of the device. To ensure that the large data volumes can be controlled and to prepare them for the development of apps a powerful middleware is required. Such middleware solutions for the platform Tango are, for example, provided by the company Scandy based in New Orleans, USA. The middleware uses the original raw data supplied by the sensors and converts this data into suitable data models that are then available for further use.
Gestures and faces
Scandy fuels the app development because access to 3D data is much simplified. The developers themselves do not have to be experts in 3D image processing because the middleware generates finished objects, e.g. a three-dimensional grid model. Two standard methods are used: SLAM (Simultaneous Localization and Mapping) and Royale developed by pmdtechnologies. With the solutions from Scandy app developers can easily create their AR apps. A typical application is recognition of a hand. The 3D model of the hand is abstracted to such a point that its movements can be evaluated quickly and reliably. So 3D gesture control can be implemented without problem. Even face recognition, e.g. for authentication, is possible without the need for detailed knowledge of 3D image processing. This new technology also allows for industrial robotics applications. A powerful middleware as a basis for development is also necessary.
Threedimensional detection and analysis with the O3D sensor by ifm. Do you want to learn more about the O3D?
ifm is a manufacturer of sensors, communication and control systems. After many years of intensive cooperation with our customers we have established ourselves in the market as service-oriented sensor specialists and today we are represented by more than 6,700 employees in over 70 countries worldwide.
© phonlamaiphoto – stock.adobe.com