Start a new topic

Use object tracking for orientation from inside the room

Can i make a photos from inside the room for build wto and then use it for place virtual objects inside the room on the walls and on any objects and move inside with phone? Can it work or i use another method for it or no way to do it?


question 2:

Can i use image recognize and objects detection together?


Hi,


Could you please share the photo's with us, so we can have a deeper look if the room and the images are suitable. You can send them to forum [at ] wikitude.com.


Thx and greetings

Nicola

In the web studio not a problem to get good result for build one wall but if i trying to upload photos of room that contains severals walls the web studio can't detect corners and build all photos like one wall.

Hi,


Could you share the images you made? In general it's advisable to take images from different angles and distances. Please also note that especially if the walls are white and there is not much structure in the room otherwise (e.g. furniture, pictures on the wall itself), the ObjectTracker might not be in the suitable quality for recognition and tracking.


Thx and greetings

Nicola

Thanks for answer. I'm tried for it "object tracking" and made a photo from difference angles inside the room and get failed with build good markers in studio around center of the room.
My question:

Do you have any advice how make a photo from inside the room? In tutorial i found only advice relate objects and large builder from one side.I can't understand should a move in the room while make a photo or stay in center (make photos from one position).
Can i build 4 difference objects for every wall of the room and detect one of the wall with your sdk and switch between them. Maybe another advice?

Hi,


You can work with our scene recognition to being able to recognize rooms. Further details can be found here. Please also have a look at our Instant / Markerless Tracking feature, which from your description sounds more suitable, depending where you want to place the virtual object and what you want to scan.


Question 2: this is currently not supported by the SDK, but we have it on our list. I can't unfortunately say anything to timelines and if / when something like this is coming.


Thx and greetings

Nicola

Login or Signup to post a comment