Unlike Image and Object Recognition, which rely on pre-mapped targets to trigger the display of digitally augmented elements, Instant Tracking is markerless. So, instead of requiring a mark, it tracks features of the physical environment itself to overlay AR content. SLAM based Instant Tracking is, therefore, highly dependent on the characteristics of the physical scene in which the AR experiences is taking place.
After discussing Image Target guidelines, in this post we will briefly talk about how Instant Tracking works, sharing best use practices and the characteristics that make up for a good tracking environment.
Instant Tracking: best use practices and environment guidelines
The ability to track arbitrary environments without the need of a marker is a trait of the Instant Tracking algorithm that enables very specific use cases. A classic application is furniture product placement as implemented in this 3D Model on Plane example. But, how does it work?
The Instant Tracking algorithm works in two distinct states:
- Initialization State: the origin of the tracking procedure is defined by pointing the device and aligning an indicator. User must actively confirm when the alignment is satisfactory before transitioning to the tracking state.
- Tracking State: in this state the environment is being tracked continuously, allowing augmentations to be placed within the scene.
For best results during the initialization state a good tracking environment must be used. Since instant tracking creates a point cloud of the scene, it relies on detecting capturable features. For that reason, an ideal scene is one that contains distinguishable elements and good lighting. A structured floor and/or carpet work best to detect ground planes, while clean surfaces might be a problem.
To get a good map of the scene, different viewing angles and perspectives are necessary. For best results take a step back after starting initialisation and move left and right around the scene.
Typically, one can walk 90 degrees to the left and 90 degrees to the right of a 3D model placed in the scene (180 degrees). However, there are cases in which a full 360 degree rotation is possible. It will depend on the trackability of the environment and the performance of the device being used.
Learn more about setting up a ideal sample scene for Instant Tracking with the Wikitude SDK in Unity by watching this Unity video tutorial. Following these guidelines should result in great Instant Tracking results. For maximum performance update your Wikitude SDK to the latest available version and make sure to use supported devices during the AR experience.
To continue reading the Wikitude AR Guide Series, access the Image Target installment. For more AR information, access the Wikitude Forum to browse through various AR topics discussed by active developers worldwide. Should you have any further questions, please contact firstname.lastname@example.org for extra support.