0
$\begingroup$

Say we have ten cameras surrounding an object. The object changes its color with a certain frequency (or any information other than color, such as a text on it). The point is if all cameras are triggered simultaneously, they will capture their frame such that the object in all those frames has the same color in it (or the same text or information).

For example, if the text on the object changes with a frequency of 1 Hz, it is obviously possible that we trigger the cameras at the same time and see the same text in all frames. However, if the frequency of the change of that text on the object changes, my question is to what realistic extent can we obtain the same information/text from all frames.

P.S., An additional question would be if this possiblity would hold if the trigger is made wireless (assuming all the cameras have the same distance from the wireless transmitter).

---edit:

All cameras look at a set of LEDs (3 or 4 ones) which are very small but their intensity can be controlled by myself so exposure time can be small (should be very small). But the LEDs move/vibrate and my goal is that the frames basically are from the same position instance. My other intuitive example is if a text is changing with high frequency and all frames have the same text if taken simoultaneously.

Improve this question
$\endgroup$
0

3 Answers 3

Reset to default
1
$\begingroup$

TLDR; it really depends on the camera protocol you are using

DISCLAIMER: My information maybe a bit out of date (last time I made an application was about in 2017 --5 years ago-- and this is an every changing subject).

IT is possible to trigger cameras using a pulse generator signal or another signal from a computer. Most professional cameras have a specific interface. I've looked at the following:

  • GigE (over network cables)
  • USB3.0
  • CoaxPress (CXP-12)
  • CameraLink

(but there are more 5GigE (an evoloution of GiGE) and BCON). You can trigger the camera's through that interface and I have been able to get a trigger several cameras close to or lower than a millisec (0.001 sec), but I am sure that you may be able to get lower. (Additionally professional cameras usually have a trigger channel that you can fine tune with sub micro sec precision using a pulse generator.)

Given that the above software interfaces give you access to all configuration options including exposure time and gain etc, I am certain it is pretty much possible to capture the same image in a time window of at least within a few ms.

Another part that may be important in your case (Depending on the application), is the type of sensor ie. CCD or CMOS. CCD's have typically have a rolling shutter (i.e. each row is taken in turn) while CMOS have a global shutter. The result might be the following for some exposure times.

enter image description here Figure : Rolling shutter of a CCD vs Global shutter of CMOS (Right) (source:Oxford instruments)

As you may guess, if the exposure time is long enough, you with the ccd you might see the transition happening in the image, while in the CMOS (with a similar exposure time) you should see the image in with all the colour information.

Improve this answer
$\endgroup$
7
  • $\begingroup$ You are right about the advantage of usage of global shutter. like the picture you have also, if my target object fibrates, I want all the frames to capture a frame from the same state of the vibration/position. So the exposure time should be as small as possible depending on the vibration frequency and image distortion like that of CCD should be avoided and only if all cameras get the frame simoultaneous, then they have all a frame from the same position of the object. $\endgroup$
    – Alejandro
    Commented Jan 15, 2023 at 16:29
  • $\begingroup$ However, what I mainly look for is a wireless method to achieve this simoultaneous trigerr of all cameras to suddenly make an acquisition. $\endgroup$
    – Alejandro
    Commented Jan 15, 2023 at 16:29
  • $\begingroup$ the wireless method inherently has more lag compared to a wired communication method. Although in the recent years with WIFI improvements and/or 5G you might be able to achieve sub second synchronization. However, I still doubt that it would be in the order of ms. $\endgroup$
    – NMech
    Commented Jan 15, 2023 at 17:21
  • $\begingroup$ 5G and wifi may allow longer distances and lag. but my work space is less than 5 meters. so I am hoping for such work space there might a more optimized solution that allows simoultaneous trigger $\endgroup$
    – Alejandro
    Commented Jan 17, 2023 at 21:15
  • 1
    $\begingroup$ I offered 5G and wifi because you explicitly mentioned wireless, and those are standards that you may find in cameras. I don't know what other wireless methods for triggering you have in mind. Also, it would be helpful if you mentioned what are the requirements for your application and the available hardware (or budget) that you have. e.g. if you have some smartphones that are want to use or having IP cameras, or and dedicated cameras the solutions are totally different. $\endgroup$
    – NMech
    Commented Jan 18, 2023 at 6:39
1
$\begingroup$

Yes, of course it should be possible to trigger all the cameras simultaneously if they are all done in parallel rather than serially or multiplexed.

If triggering wirelessly you may have problems with channel interference unless you use one trigger and all receivers on the same channel.

You haven't given any detail about the camera types and where the frame grabbing is done. This too might be a bottleneck and cause the data to come in with a time offset depending on whether the camera has a buffer or not.

Improve this answer
$\endgroup$
5
  • $\begingroup$ what do you mean about channel interfaces? Do you have commercial examples also? All I want from the camera is actually to have an image of 4 LED (that I also design) and put 20cm away to have a very accurate image of their position in the 2D image. Like imagine your target object vibrates and you want all your cameras to get a frame from the very same position of the object during its vibration (so global shutter is needed and most likely Cmos camera is better) and exposure time should be very small. At least I can improve the intensity of the LEDs and they're only 20cm away. $\endgroup$
    – Alejandro
    Commented Jan 15, 2023 at 16:22
  • $\begingroup$ The question is also how perfact can they simoultaneously be triggerd with the current state of the art, especially wireless. like what may be the limit now or what time deviation/variation may be expected between the frames (of course speed of light and distance of cameras for example play a role here but we neglect them).. $\endgroup$
    – Alejandro
    Commented Jan 15, 2023 at 16:25
  • 1
    $\begingroup$ "... channel interfaces?" I said "channel interference". That may apply if you had multiple transmitters. The deviation / variation would be specified in the transmitter / receiver data sheet. One way to address that if it were essential is to have a GPS receiver in each device to synchronise the clocks and then send a command to trigger the cameras at some specific time in the future - possibly only a few milliseconds into the future. $\endgroup$
    – Transistor
    Commented Jan 15, 2023 at 18:39
  • $\begingroup$ hmmm using a GPS that allows synchronization of the clocks in different circuits, do you have an estimate how simoultaneous the trigger can then be? For example if we have a clock in each different device, how perfectly can we actually synchronize them? $\endgroup$
    – Alejandro
    Commented Jan 17, 2023 at 21:11
  • $\begingroup$ @azerila, no, I have never built anything with GPS. $\endgroup$
    – Transistor
    Commented Jan 17, 2023 at 23:27
1
$\begingroup$

Regardless of the data interface, many cameras offer hardware triggering (sometimes software configured). Once you've got a hardware trigger you can wire whatever you like up to it, crucially connected in parallel.

I've used hardware triggering down to microsecond-scale repeatability and jitter on Basler Ace cameras, for example. I would expect several cameras of the same model to trigger at the same speed, and with some cameras you can calibrate the trigger delay. Some models in the range go down to 1mus exposure times, all or most have hardware triggering. There are rival makes, but I can't recall the other one I used with similar specs, and I'm working on slower cameras these days.

I've also built hardware triggers for DSLRs, but never tested their timing accuracy. I'd expect millisecond scale or quite likely far better, but ~millisecond exposure times are as fast as you'll get with those.

For wireless triggering, there may be off-the-shelf solutions, but in a controlled environment I'd look at triggering on a simple infrared pulse with a phototransitor-based receiver (possibly filtered optically) per camera and one transmitter with several LEDs- just be sure not to pick up the trigger pulse on the cameras. Timing accuracy form propagation delays isn't a problem with µs-scale timing as that would be hundred of metres at light speed

Improve this answer
$\endgroup$
2
  • $\begingroup$ I actually look for sub-microsecond exposure time and global shutter. Ot at least a few microseconds $\endgroup$
    – Alejandro
    Commented Jan 23, 2023 at 18:40
  • $\begingroup$ That desired spec should probably be in the question, but some of the Basler ones I mentioned have a good chance of coming close (down to 1µs exposure), hardware triggering. $\endgroup$
    – Chris H
    Commented Jan 23, 2023 at 21:29

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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