I'm trying to create an application where a camera is looking at an object continuously. The object will receive small changes, imagine small parts of it being blocked for the camera. I would like the camera to automatically detect this.
I have this working when i shake the camera and it 'rescans' the object and sees it's changed, but I can't get it to work without shaking/moving away from object.
Is there a feature or possibility to make this work?
Good morning Simon,
the only way I see this working with our SDK is if the change of the object is significant enough for tracking to be lost and re-found using a different target.
You could, of course, lose the tracking programatically by disabling the tracker when the change occurs, but that would probably not be compatible with you requirement of the change being detected automatically, would it?
Thanks for the response Daniel,
Unfortunately the changes we're talking about are things like adding ram or a harddrive in a laptop, so the changes are pretty minor and won't cause the object to be lost.
As for disabling the tracker, I can't get this to work either since the changes are never recognized in the first place without looking away from the object.
The reason I had this question is because I'm working on an internship and I'm basically completely done except for this automatic change detection. But since it seems to not be possible right now, is there any talk about this feature being added in the future?
Thanks again :)
I don't think we have anything in that regard planned currently. To be honest, I don't really know what this feature would look like.
Maybe you could, with the current version, have different targets of your laptop. Each focussing on a different part of it such that adding a component would be a significant change for the target focussing on this particular component. For example: have a target that has most/all of its feature points in the area where the ram will be inserted and have another target that has its feature points in the same area, but with the ram already present. Maybe you could get the former to lose tracking when the ram is being added and have the latter take over afterwards. I'm not sure how well this would work practically though.
Thanks for the response again, I can use that in my report for sure.
The example you've given is pretty much exactly what I got. One pic of the laptop without the ram placed in and one with the ram placed in. I could possibly zoom in on the ram or something maybe that could work somehow so I might try that. However this is meant to eventually be implemented on an assembly line kinda deal and you can imagine it not being very practical that way.
Thanks for your help anyway tho :). I'll close the topic after this message.