topic Hi Sir: in Software Archive

Realsense 3D Linux support

TChio — Tue, 17 Nov 2015 05:37:11 GMT

Hi Sir:

Could you help below some question for deep and IR question ?

Environment OS : ubuntu

now we could read r200的 RGB, IR, depth info，but the IR & depth image is different with RGB。So only show green screen (as below picture.)

1. IR & depth’s format could use on Linux? How to check this and add it on Linux ?

2. How to use the corresponding format, read IR and depth of information and this method is applicable to Linux

3. how to Verify whether there is to read information correct? Or simply just to judge from the eye image?

Regards

Tab Chiou

Currently RealSense SDK only

Xusheng_L_Intel — Tue, 17 Nov 2015 16:49:16 GMT

Currently RealSense SDK only supports windows 8.1 and windows 10. Thanks!

Hi David:

TChio — Wed, 18 Nov 2015 02:26:28 GMT

Hi David:

yes ,i know but we not SDK support. we just want to know how to read the IR, depth format.

could you help this ?

Regards

Tab Chiou

Someone got it to work on

Henning_J_ — Wed, 18 Nov 2015 09:29:30 GMT

Someone got it to work on Linux (I think F-200 only): https://github.com/teknotus/depthview

Demo Video and more details linked from the Github page.

There's also a ROS package for the R-200: https://github.com/PercATI/RealSense_ROS

Hi Sir:

TChio — Mon, 23 Nov 2015 03:43:08 GMT

Hi Sir:

one more question.

According r200 data sheet, the depth stream format is "Z16_2_1", and its description is "16 bits, 2 pixels packed into 1 32 bit word".
Can you please illustrate what "2 pixels packed into 1 32 bit word" is, because the format is not common in video formats.
Besides, please also provide the F200 data sheet to us. I want to compare r200's format with f200's.

Well, there isn't many

samontab — Mon, 23 Nov 2015 04:16:20 GMT

Well, there isn't many standards for 3D video, so no wonder it is not standard.

It just says that they are using a 32bit pixel to store two 16bit pixels. You could use any standard that uses 32bit pixels to read a temporary value, and internally treat it as two 16bit pixels.