- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi,
I am wondering how to make the most of the R200 camera to be able to detect ranges at about 3-4 metres (or above). I realise now that the ivCam parameters do not affect the R200 model. I am not interested in gesture tracking, only scanning a space (which I know was not the intention for the camera). I find that walls can be very patchy and I do not manage to get enough information. Is there anything I can do to improve things? I'm using the 628x468 depth resolution.
From searching the forums it seems that I can change DS_MIN_MAX. Are there other parameters I can tune to help in this situation? (e.g. with the ivCam code I was interested in changing laser strength and filtering).
Thanks
Link Copied
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Changing DS_MIN_MAX wont help you. That just changes to range of the values returned, not the quality of those values.
But you can change frame rate, exposure, and gain of the IR cameras. That will affect the depth values. For some details, see the R200 Design Guidelines (page 9 ff.). The corresponding SDK methods are PXCCapture::Device::SetDSLeftRightExposure() and PXCCapture::Device::SetDSLeftRightGain().
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Warning to all: do not use 628*468 depth resolution on the rear camera!
I've just spent the best part of a day trying to figure out why my depth mapping worked perfectly on the F200 but was giving really weird values using the R200. Turns out the 628*468 depth image is a jumbled up mess; switching it to 480*360 solved everything immediately!
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
James B. wrote:
Warning to all: do not use 628*468 depth resolution on the rear camera!
I've just spent the best part of a day trying to figure out why my depth mapping worked perfectly on the F200 but was giving really weird values using the R200. Turns out the 628*468 depth image is a jumbled up mess; switching it to 480*360 solved everything immediately!
Can you show an example of the image you get with 628x468? I have been using that resolution for a while without many problems.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
I wasn't generating an image; I was using projection.MapColorToDepth and projection.QueryVertices to get the 3D positions of points mapped from the colour image. Using the 628*468 resolution I was getting crazy values, saying I was metres away if it gave me anything at all. As soon as I switched the resolution it worked exactly as expected and as it has always done on the F200.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
how to acquire a good quality in depth image?
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hey guys,
I am working in improving the depth image capture from video from R200 in Linux.
I am looking for the equivalent of: PXCCapture::Device::SetDSLeftRightExposure() for LIBREALSENSE.
thanks
![](/skins/images/7746D4550810124E96DD2FD9957552B1/responsive_peak/images/icon_anonymous_message.png)
- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page