How to change Kinect visual layers

[kleinash]

Hi,

Is there a way to put an object out of the view of the Kinect? i.e. I would like my Kinect vision sensor to just see the object and not the Barrett hand or the floor, is that possible? I have tried turning on different visual layers for the object and Kinect. I am dealing with it by just moving everything off the ground, and the hand out of the way for a moment.

Regards
Ashley

[coppelia]

Hello Ashley,

you can specify which object is visible to vision sensors, and which is not. To be visible by vision sensors, your object needs to be renderable. Vision sensors ignore the layer information of an object (only cameras take that information into account). In next release you will also be able to specify for each object which camera/vision sensor can see it.

Cheers

[kleinash]

Hi,

I really appreciate how quickly you answer..

The renderable box under object special properties is already unchecked (for the Kinect), is there perhaps somewhere else I need to be looking? I was able to uncheck for the object and it worked as expected..

Is it also possible to make the floor unrenderable?

Off topic but I was not able to reply to this thread, something about being too soon to post again..

Regards
Ashley

[coppelia]

Ashley,

you have to be careful, some objects are hidden in a different layer, but if they have the renderable checkbox checked, then they will be visible when seen from a vision sensor.
Cameras and vision sensors operate differently.

When looking at the floor, you will see that it is composed by several objects. You will have to modify the properties of each of those objects, depending on what you want to do.

Cheers

[kleinash]

Hi,

I see, thank you.. For the moment - to avoid having to perform segmentation - I am going to just work with the object in the air, take the first image as being separate and then introduce the barrett.

Thank you for your time.
Ashley