Start a new topic

dynamic trackable2DObject & Image/video generation

dynamic trackable2DObject & Image/video generation


Hey there

I'm currently stuck with a problem: Is it possible to dynamically add trackable2dobjects and resources?

I want to use the cloudrecognition with metadata. The Idea is as following: the tracker will recognize an object based on the images in the cloud. I get the metadata and in there is the information (video or image) and url. Now I'd like to add a new trackable2dobject, give the object the resource that is in the metainformation and display it.

This means that I need to create trackable2dobjects since I never know who many trackable2dobjects are visible at a time. The examples always work with one or 2 hardcoded trackable2dobjects, is there a way to add and manage them dynamically?

How dos the onRecognition function work, does it call the response every time a target is found for every target on an image? Does it call every single frame all the targets or only targets that were newely recognized?

Hope you can help me out

 

Hi Dietmar,
Of course you can add them dynamically. If you know, based on the metadata, which objects you need to create, you can create them. To create a Trackable2DObject, you only need to know to which tracker it belongs (The CloudTracker) and which drawables should be shown on the target. You can also add new drawables to a existing Trackable2DObject.

onRecognition is called everytime the server could process the given camera frame, so it's not called every frame. A new server response can eiter be positive (a target was recognized) or negative (no target image was found in the camera frame).

 

Was this helpful for you?

Best regards

Andreas

ok thanks for that!

I wonder if the same is possible with the ClientTracker since there is no onRecognition event. Can you point me into the right direction or is it even possible to dynamically create resources and trackable2dobjects with the clientrecognition? Would be needed if I'd load the newest wtc file before I'd start the clientTracker.

?dit:

More importantly I'd like to know if it is possible to have multiple targets on one screen with the cloud Tracker. I have a working solution with generated trackable2DObjects but there is only 1 active at a time. Looks like the on recognition only gets called if the last recognized tracker is outside the Screen. If I scan 2 objects after another I only see an overlay over one of the two and it will switch as soon as the visible overlay is outside the window.

Hi Dietmar,
You can also dynamically create Trackable2DObjects that refer to client tracker. Client and cloud tracker behave the same in this regard. You can load/download different .wtc files at runtime using a client tracker and use them to create trackable objects.

Currently both tracking solutions (client and cloud) do not support multiple simultaneous recognitions. You can only have one active tracker which will search for only one target at the same time.

Best regards

Andreas
Login or Signup to post a comment