3
\$\begingroup\$

This is my first post in this forum! So Point me if anything is not clear!

I want to know how TOF camera works? So I've done Google & read some nice explanation about How Phase difference between Transmitted & Received signals can be used to calculate the Depth of an Object from Camera with good example.

What i did understand:

  1. How a light source(IR) can be modulated & demodulated ? Answer
  2. How to filter the reflected IR signal having the same modulation frequency? Answer
  3. Depth Can be measured by Calculating the Phase change of the Transmitted signal.

What i cannot understand:

  • How phase change of a continuous square wave signal can be used to Measure the depth?

Update: Here is a Nice Explanation about this.

Note: I'm talking about TI & Mesa Kind of TOF Camera's using IR Led

\$\endgroup\$
  • 1
    \$\begingroup\$ Time of flight should not be using phase shift, that sounds more like Continuous Wave Phase Shift style of laser range finding. The IR camera would be pulsing an IR light source, and the camera pixels would accumulate a charge on their plates. The assumption is the less charge they accumulate the further away the target is. I think! \$\endgroup\$ – KyranF Dec 17 '14 at 14:10
  • \$\begingroup\$ @KyranF I guess this is not the case in TOF! Don't you think so? Check my Links \$\endgroup\$ – Balaji R Dec 18 '14 at 5:34
  • \$\begingroup\$ Traditional time of flight, especially with lasers (not CCD/CMOS cameras) does not operate this way. It seems the terminology is a little messed up, and they HAVE to modulate the IR emissions otherwise the whole thing would never work in sunlight or near strong sources of IR (heat, for example). The modulated output means they can have filtered input and get more reliable results, including the inherent ability to check for phase shift. Phase shift, depending on the frequencies used, can be very inaccurate or if you want accurate, it's "range"/wavelength can overlap and you get ambiguous nums. \$\endgroup\$ – KyranF Dec 18 '14 at 13:42
4
\$\begingroup\$

The Kinect camera is NOT a TOF (time-of-flight) camera. Instead, it projects "structured light" (a pseudorandom pattern created by an IR laser and a hologram) into the scene, and analyzes the image it gets back through a second (IR) camera in order to determine depth.

\$\endgroup\$
  • \$\begingroup\$ Thank you for your answer @Dave. I was wrong! i was asking about TI TOF Camera's using IR signals. I've edited my question. \$\endgroup\$ – Balaji R Dec 17 '14 at 16:23
  • \$\begingroup\$ I think the old one was like this but new one for xbone is TOF? gamasutra.com/blogs/DanielLau/20131127/205820/…. Not being a dick :). I was just interested to learn more about them. \$\endgroup\$ – Some Hardware Guy Dec 17 '14 at 17:54
3
\$\begingroup\$

The kinnect uses an array of ir LEDs, actually it's a single chip array I think. This outputs a grid pattern. They bought two companies with this tof tech. On seems to work by modulating the brightness of these LEDs with an rf carrier. So maybe a dc voltage to keep LEDs lit, then a carrier is added in on top of that.

Then they use a high, 160fps, image sensor to record the scene. After that with some processing of the image they can measure the phase of the incoming led light reflected off of the subject being measured vs the phase of the transmitted RF carrier. The phase difference is from travel time and from that you get distance.

Another approach appears to be that same field of LEDs but with an electronic or mechanical shutter running in front if the array. Here some the amount of light returned will be blocked by the shutter. How much depends on it's time of arrival and from that they can infer distance traveled.

Wiki has a good explanation of the different types.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.