Advice for ZR300 multicam setup / settings

[DKors]

I'm currently attempting to do something similar to a real-time 3D scanner using a multicam setup.

I bought the ZR300 realsense due to this description in the datasheet: "In general, the ZR300 is not subject to infrared projector interference from multiple cameras in use simultaneously." (https://click.intel.com/media/ZR300-Product-Datasheet-Public-002.pdf source)

Now I have been running a small test setup using the multicam-cpp example bundled with librealsense with 2xZR300 and a 1xSR300 connected. I put a small video up of what I found below:

The two ZR300's are the left and middle windows, while the SR300 is the right-most windows. Looking at the depth image SR300 has a very nice clean depth image with almost no missing data (black pixels). After putting the SR300 away the two ZR300's still have more noise and it becomes obvious (look at timestamp 0:35 to 1:40 in the video) that the areas which both IR projectors hit are blocking the depth measurements from for both ZR300 (creating black pixels). I find this a little odd thinking of the quote from the data sheet, but maybe I'm misreading "interference"?

I read in the data sheet that the active projector is only there to help in scenarios where texture is missing (e.g. measuring depth on a unicolored wall). Would it be possible to turn off the projectors and only rely on the stereo vision based depth? Do you have other suggestions for optimizations that I could do to improve on this multisensor setup?

Thank you.

|---|---|

| Camera Model | ZR300 |

| Firmware Version | No idea |

| Operating System & Version | Windows 10 |

| Software | librealsense compiled in VS2017 Community |

[idata]

Hi da9e,

Thank you for your interest in the RealSense cameras.Let me investigate your issue a little bit more, I'll get back to you once I have some updates. We appreciate your patience.Regards,Pablo M.

[DKors]

Thank you for looking into this. Any recommendations for optimizing the multicam setup would be appreciated - down to what kind of materials that surfaces should be made of/look like and settings for the cameras etc.

[idata]

Hi da9e,

Unfortunately, our own experiments have shown the same results that the user sees. We are working with our engineering team to clarify this claim in the datasheet and for further guidance. Our experiments also show that at further distances, over 2 meters, there is no noticeable interference. Right now this is the information we have regarding this.Regards,Pablo M.

[idata]

Hi Da9e,

Do you have any further question regarding the ZR300 camera? If that's the case, let us know so we can try to help you.Regards,Pablo M.