Update (August 2017): Object recognition, multi-target tracking and SLAM: Track the world with SDK 7
With this post we are opening a new chapter on Wikitude’s journey towards augmenting the world! We are happy to share today the first version of the all new Wikitude 3D tracking technology. For our team, augmenting rooms, spaces, and objects around us is a natural progress after mastering augmentations on 2D surfaces. Clearly, tracking in 3D is a much more complex task as algorithms must be optimized for a variety of use cases and different conditions. With this BETA release of our 3D tracking technology, developers will be able to map areas and objects of a rather small scale and place 3D content into the scene. This is the first step of a sequence of releases Wikitude will roll out as our SLAM based 3D recognition and tracking technology evolves. The 3D tracking feature is now available as a free trial and packaged in our SDK PRO products. This feature is currently available for the SDK 5 Native APIs only.
How does Wikitude 3D tracking work?The Wikitude SDK tracks 3D scenes by identifying feature points of objects and environments. By identifying feature-rich environments, the SDK will map the scene by displaying a point cloud over the different feature points. As an example of how the Wikitude 3D tracking works in a small scene, we will use the scenario of an office table. The richer the scene is equipped with feature points, the better the mapping and tracking will be. In order to track and map the scene the following steps should be taken:
- Launch the Wikitude sample app, which is included in the Native SDK (iOS and Android) download package
- Record a tracking map by slowly moving the device from one side to the other of the scene, covering the whole area
- 3D point clouds will appear on the screen capturing key feature points of the scenario
- Save the tracking map
- Load the map in your augmented reality experience to relocalize the scene and visualize the augmentation in real time.