1

IR image of the reflective markers

Hi, I'm doing my master thesis on collaborative robotics in which I use 4 spherical markers covered with reflective tape in order to detect them with the IR camera of the Kinect. This avoids problems of noise given by the use of colored markers and RGB camera but introduces a very big limitation. I cannot retrieve depth informations of this markers!! I was wondering what is the physical explanation of this. I mean, theoretically depth informations work with TOF sensors and a reflective materials reflect back the light rays emitted, where is the problem? Thank you in advance, Ale

Xela95
  • 63
  • 7
  • didn't use kinect v2 yet, does it have problems with different materials like mirrors? What kind of reflective tape do you use? – Micka Feb 08 '20 at 15:38
  • @Micka it a simple reflective material, it has a similar effect to that of a very polished surface, it reflects almost all the IR rays that receives. Accessing the depth information for the pixels of the marker I obtain a null measure. Maybe because it reflects IR rays in other directions? – Xela95 Feb 08 '20 at 19:51
  • kinect v2 uses modulated light. I dont have enough physics background, but could it be that your special surface changes the modulation? https://www.researchgate.net/figure/Kinect-v2-uses-the-time-of-flight-method-for-depth-sensing_fig3_277637546/amp – Micka Feb 08 '20 at 22:27
  • are your markers more like cat's eye reflectors? Do you observe a similar behaviour for other materials (unpaintee metal, mirrors, ...)? – Micka Feb 08 '20 at 22:29
  • @micka First of all thank you for your time to answer me :) yes my marker has an effect like a cat's eye, but this behaviour is the same for materials like mirrors and lucid metal. The thing that I don't understand is why TOF sensor cannot compute the phase shift for rays impacting against these surfaces... – Xela95 Feb 09 '20 at 13:36
  • I don't know whether modulation is changed (then the kinect doesn't know anymore which reflection/depth belongs to which pixel. For mirrors you wouldnt see the reflection on the mirror but on a different object (IR-Laser to mirror to object to camera), maybe resulting in.multiple pixels for the same laser ray. For cat's eye I could imagine that multiple laser positions are bundled in the cat's eye center, so a single pixel would get multiple depth's. Both might confuse the kinect. But not sure whether this really is the reason... – Micka Feb 09 '20 at 16:59
  • "Speckle noise, dynamic range of light intensity, and spurious reflections are major challenges when laser scanners are used for 3D surface acquisition" https://www.hindawi.com/journals/ijo/2017/4134205/ – Micka Feb 09 '20 at 17:02

0 Answers0