I need a strobe to illuminate Arduino/RPi camera field

I have been on this forum before trying to figure how to count plankton using an Arduino or RPi system deployed from a drone. Several responses were very helpful, steering me away from a beam breaking system toward imaging with a camera. But now I need to stop the action that the camera is imaging since the tiny (~0.5 mm) plankton will be flying by in a flowing stream. I think I need VERY short duration illumination to stop the motion through the image field. I assume that LEDs flash on for too long a period to stop the movement and that I probably need a very small strobe working in conjunction with a camera shooting video. Once the video is shot I then grab an illuminated frame from the video and post process to to count the plankton. Thanks to the form I have tried the system on static plankton (preserved in lab) and it works well but now I need to work on illumination to stop the motion in the field. Any thoughts on whether there are very short flash LEDs or small strobes that could be worked into the system - I can't find any small hardware to do the job. If found can it be incorporated into the Arduino/RPi to work with the camera video?

The project: I am studying a nearly-extinct whale, the North Atlantic right whale, and we need to figure out how much plankton the whales eat and where the food is by, hopefully, using a drone to greatly improve the spacial/temporal resolution of the plankton resource around feeding whales.

I wonder why someone would ask such a question: "I assume that LEDs flash on for too long a period to stop the movement" on a hobby forum when a little research on LEDs will give you the answers.

You do not give any hint as to the color of LED you are going to use and there in lies your problem. You cannot rely on your human vision to discover how long a LED will continue to produce light after the electricity is removed.

If you are thinking of "white" light LEDS, you should research them. They are made unlike any other LEDs in that they use phosphurs to produce the "white" light, not the LEDs themselves.

If you want white light, then research for devices that fit your requirements.

Paul

I am wanting to try white light first. I am so far out of my lane and but desperately trying to find the answer - looking at keywords I could not find an answer or figure out how fast the flash could/should be. Sorry to have asked what for me (an utter newie)is an opaque question, I will try to look into LEDs for stopping fast action, thanks.

Darksea13:
I probably need a very small strobe working in conjunction with a camera shooting video.

It sounds a tall order, but at least very small strobe might imply a very weak one. I guess it depends on the frame rate of the video. You might find that continuous LEDs are the answer.

White LEDs use a phosphor which glows for some time after activation, so are not particularly suitable for a strobe.

RGB LEDs should respond - and go dark - much more rapidly.

Thanks for your thoughts Nick and Paul. The problem is that the flash of light must be so fast as to stop the rapidly moving and very small plankton organisms that will be traveling in a water flow at about 10 cm/sec and if the light, as in the case of a strobe, has a bright enough yet very short duration flash the plankton will be frozen in the frame of the video. A bright red LED made to display a short flash might be what I am looking for but I still cannot find specs for the minimum duration of the flash.... assuming that I can get my hands on the code needed to create the shortest flast possible.

A high voltage strobe, not a typical camera flash, can flash pretty rapidly, that may still require you to sensibly define "video" and properly consider what your real needs are..

Even taking Paul's comments on board, I imagine Arduino would be quite capable of flashing the appropriate LED at full power at any "video" speed, and it wouldn't surprise me if the lights you see on video cameras do exactly that in order to save power.

You also don't seem to have considered using a steady light and controlling the exposure in the camera. My ancient webcam will shoot at 90fps, which implies a pretty fast shutter speed. I imagine you can do that with a phone these days.

I'm not sure this is an Arduino problem, but you do need to do some proper research.

Again thank you Nick. I have looked into the possibility of using a fast frame rate but it appears that I need around 1/10,000 second or less flash duration to capture images of <0.5 mm organisms, particularly their very fine detail, in water flowing at 10 cm/sec. I have actually tried frame rates up to 60 fps and the image quality is poor to very poor because of the movement of the organisms while the frame is being exposed, so with the suggestion that I need as little as 1/10,000 second flash duration I have nevertheless continued to look at LEDs.
While I cant get my hands on specs as to the shortest duration what I am hearing here and on the internet is that an LED won’t reach anywhere near that short duration. I have been looking at inexpensive traditional flash units for DSLRs and they have power control (and at low power and illumination their specs say they can get down to 1/25,000 sec). Still I would really like to go with LEDs and all that I see on this forum makes me keep researching - still I can’t find duration specs of the shortest durations for HBLED. Any further thoughts are most welcome - thank you Nick.

Darksea13:
I have been looking at inexpensive traditional flash units for DSLRs and they have power control (and at low power and illumination their specs say they can get down to 1/25,000 sec).

Indeed, but DSLR is not video, hence my comment about your need to define video - or consider if you need it all. I would be surprised if any traditional flash can deliver more than 5 fps or so and would likely be useless if they did, and anything faster would use continuous lighting.

Nick: again thanks for thinking through my attempts to build a plankton imaging system. I guess I have not been clear - this is how I am imagining the system working (and what I have tested already using a flashing LED BUT not using a very short duration flash): the video (Arducam video)films the cell while occasionally the flash fires - in this system there is no syncing the video frames with the flash. Every several video frames the flash is captured. In my test there was no flowing water and the plankton was preserved, hence a static system for the test. On post processing analysis I selected those frames that were illuminated and I discarded those that were black. In this system the video runs and the frame rate is relatively unimportant because it is not used to stop the action (even though, in the tests I ran there was no movement and so no need to stop a static sample). The results were gratifying - the illuminated frames that I selected showed detailed images of preserved plankton. With that success I now confront the last big hurdle, how to stop the action in a real flowing water system. I believe then that a very short bright flash, probably much shorter than an LED flashes, will capture the plankton in the frames that are illuminated BUT being a very short flash (like a regular strobe) the rapid motion of objects in the field of each illuminated frame will capture crisp images of plankton. The DSLR flash unit would only provide an occasional very short flash (a DSLR itself would not be used, only a hot shoe flash with power selection to shorten the flash duration) to illuminate occasional frames captured by the Arducam video stream (eventually to be stamped with data from GPS, Time, temperature and salinity). What remains a central question is whether there is a LED that can, perhaps with programming, be made to produce the very short flash needed to stop action. I just cannot find data on the duration of LED flashes or any data on how programming might be implemented to shorten the flash. Thanks again for thinking this through....

I suspect you can't find any information on LED light/off time because all LED functions are basically analog. There is no consistency From device to device for such a parameter even from LEDs from the same batch.

I suggest you set up a test circuit to turn a LED on and off using an Arduino and monitor the light output using a sensitive diode or other light sensitive component, and watch the response with a digital oscilloscope. Then pick a few LEDs that function properly in your desired time frame.

Paul

What about a constant light source, and a shutter designed to expose the short bursts you need... e.g a spinning slotted disk ?
Depending if this suitable, there would be several methods to sync the assembly.

Channelling Nipkow! :grinning:

Hi,
I think you need to look at this, HIGH SPEED LED FLASH, claimed flash duration down to 0.5us.
It includes parts and description and uses 328P.
It may need scaling down for your application, which would mean lower power requirements.

All I did was google

high speed photographic led strobe light

Tom... :slight_smile: