In AR applications, it is essential to recognize and track target objects in the scene to enable AR effects based on those objects. Tetras.AI's marker recognition and tracking technology can quickly and accurately estimate the 6DoF position and orientation of plane image markers and 3D objects in the environment based on RGB image data, and can track the relative position and orientation of the markers in real time. Users can detect and track the location of the image or object in the real environment in real time through mobile terminals to achieve consistency between AR effects and the real space. This technology provides basic AR capability support for smart phones, urban cultural tourism, commercial real estate, cultural creation and other scenarios. Our marker recognition and tracking technology supports a variety of terminals such as mobile phones, mobile operating systems such as Android and iOS, and lightweight platforms such as H5 and mini programs. It supports multi-target recognition and positioning under the framework of terminal-cloud integration.
Android 8 and above, iOS 10 and above, WeChat mini programs,etc
Our recognition and tracking algorithm can support static and dynamic markers, with the algorithm supporting native mobile phone and web platforms at the same time.
1. Takes about 100ms on average to recognize images on mobile terminals, that supports 1/20 image proportion recognition, recognize a max. 45-degree parallax angle recognition and perform multi-image recognition and tracking at millisecond-level tracking speeds on mobile terminals.
2. The 3D object recognition and tracking algorithm supports CAD model-based recognition and tracking. Taking about 60ms to recognize objects on the cloud, it can handle recognition in complex illumination environments and backgrounds, and has millisecond-level tracking speeds on mobile terminals.
Lightweight image/object recognition and tracking technology designed for mobile terminals.
Our solution incorporates a fast and robust algorithm for image and object recognition and positioning. It supports image recognition and positioning with large viewing angles as well as smaller ratios, and can accurately recognize objects with weak textures.
The technology also offers efficient and robust image and object tracking, ensuring stable tracking at the millisecond level. It supports the tracking of both static and dynamic objects, enhancing the overall tracking performance.