Recognizing environment limitations in an AR scene

The following list identifies additional limitations that you should consider when performing an environment recognition and building an AR scene.
  • Objects
    The AR engine has difficulties recognizing:
    • Transparent objects, such as glass and windows
    • Highly reflective objects that are made of shiny materials or metallic surfaces
    • Black objects

    Do not spend much time scanning these objects, because the AR engine cannot often recognize stable feature points on them.

  • Occlusions (obstructions)

    Before you begin to scan an AR scene, look for objects that may occlude or obstruct the line of sight to a targeted device. This may cause the AR system to not identify the device correctly. If an occluded device is important for the AR experience, make sure you scan those areas carefully so the AR system can correctly identify the device and its location within the AR scene.

  • Light conditions

    Light is an essential element for the AR system because it is based on what the camera is able to detect. Make sure that the environment has ample light to ensure the camera can provide the AR system with the best image quality.

    Most mobile phone cameras automatically adjust to provide a good response in dark conditions. However, dark conditions can decrease the image quality and affect how the AR system detects the device within the AR scene.

  • Patterns

    Feature points are unique regions or spots in the environment. Areas with symmetries or repetitive patterns are difficult for the AR system to detect and process feature points.