deep-sea autonomous shooting - Script Writing - CHDK Forum
« previous next »
Pages: [1] 2   Go Down

deep-sea autonomous shooting

  • 12 Replies
  • 7119 Views
*

Offline glowgoose

  • *
  • 4
deep-sea autonomous shooting
« on: 02 / November / 2009, 19:58:39 »
Advertisements
Hello folks,

I am a marine biologist, contemplating the use of CHDK to program a Canon G9 camera to shoot autonomously at ocean depths of 2 km and deeper. I would greatly appreciate an advice on appropriate scripts:

The first application is time lapse macro shooting of giant amoebas that we recently discovered (http://news.bbc.co.uk/2/hi/science/nature/7739703.stm). Interval between shots 5-15 minutes, time to first shot 2 hours (that's how long it takes for the sub to go down). It must trigger the external, but not internal, flash each time.

The second application is shooting in response to an external event. Perhaps I could use motion-detector script, if only it will work with a far-red (invisible for ocean creatures) light source as a background illumination? The shots would then be taken with a flash.

Generally, I welcome all the crazy ideas along these lines.

One other thing: can the script / shooting sequence be restarted from a connected computer? Then I could make a port on the housing and will not have to take the camera out of the housing after each dive, which would make the whole enterprise much more fail-proof.

cheers

Misha
http://www.bio.utexas.edu/research/matz_lab
Logged
*

Offline reyalp

  • ******
  • 14128
Re: deep-sea autonomous shooting
« Reply #1 on: 02 / November / 2009, 21:16:28 »
Welcome, this sounds like a very cool project.

Quote
Interval between shots 5-15 minutes, time to first shot 2 hours (that's how long it takes for the sub to go down). It must trigger the external, but not internal, flash each time.
Depending how long and how many shots you intend to run total, you may need an external power supply. CHDK can't power up the camera at a specific time, so it would be sitting idling for the first two hours unless you added some external mechanism/electronics.

Quote
The second application is shooting in response to an external event. Perhaps I could use motion-detector script, if only it will work with a far-red (invisible for ocean creatures) light source as a background illumination? The shots would then be taken with a flash.
All powershots have an IR filter. Some people have removed them, but on most cameras it is a fairly delicate process. ISTR there are people who do this commercially for some models as well. Not sure how far into the IR they go.  You would also end up with the IR in your final image.
Quote
One other thing: can the script / shooting sequence be restarted from a connected computer? Then I could make a port on the housing and will not have to take the camera out of the housing after each dive, which would make the whole enterprise much more fail-proof.
Not currently, but there is active work to support this kind of thing:
http://chdk.setepontos.com/index.php/topic,4338.0.html

Assuming there isn't too much metal in the way, you might also consider eyefi http://www.eye.fi/
Logged
Don't forget what the H stands for.
*

Offline Microfunguy

  • ****
  • 4037
    • StereoData Maker
Re: deep-sea autonomous shooting
« Reply #2 on: 03 / November / 2009, 05:13:33 »
Quote from: glowgoose on 02 / November / 2009, 19:58:39
Interval between shots 5-15 minutes

Bearing in mind that the resulting time-lapse movie is played at 30 frames per second, over seven minutes of real time would playback in one second.

Quote
It must trigger the external, but not internal, flash each time.
The gentStereo device can do that and generate the time-lapse pulses.
It will shortly be used for aerial photography in Antartica.
Details here http://www.gentles.ltd.uk/gentstereo/options.htm

(SDM is a specialised version of the basic CHDK)
Quote
Perhaps I could use motion-detector script
Motion detection works by comparing current-average of luminance of the selected pixels in a cell with the previous average for the same cell.
If the absolute difference exceeds a preset value it is recorded as a 'hit'.
Problem is, in low contrast and/or illumination scenes quite large movement can give quite small differences.
So, it may or may not work.
Even with its inbuilt IR-cut filter, the camera will still record IR (just try pointing your TV remote at it).

Quote
can the script / shooting sequence be restarted from a connected computer?

In a sense ....
If you can generate an on/off signal of about 3.6 to 6V that can be connected to camera USB connection and script started and paused by monitoring the USB.

I have used serial port to generate the +5V signal.

To quote the SDM documentation  :-

"wait_for_switch_press

Script pauses until voltage signal present on USB V+ pin.
See the various examples above.
Signal can be from a switch, IR or wireless receiver, microcontroller, etc."

For details of other commands see here http://stereo.jpn.org/eng/sdm/ubasic.htm

For uploading your images to PC at regular intervals see here http://stereo.jpn.org/eng/sdm/upload2.htm


David
Logged
*

Offline glowgoose

  • *
  • 4
Re: deep-sea autonomous shooting
« Reply #3 on: 08 / November / 2009, 17:19:23 »
Thanks for responding, folks!
Let me try to put my most pressing questions directly:
1. can powershot G9 be programmed for time-lapse with 5-15 min interval?
2. can it be programmed to trigger an external flash without flashing its own?
3. how long will it live/shoot on its own battery? (sorry, this one is definitely not a CHDK question, but perhaps you just know this)
Logged
*

Offline Microfunguy

  • ****
  • 4037
    • StereoData Maker
Re: deep-sea autonomous shooting
« Reply #4 on: 08 / November / 2009, 17:51:15 »
Quote from: glowgoose on 08 / November / 2009, 17:19:23
1. can powershot G9 be programmed for time-lapse with 5-15 min interval?

Yes.

Quote
2. can it be programmed to trigger an external flash without flashing its own?

I am tempted to say no  ....

However, if conditions are quite dark you could set a shutter-speed of maybe one second,  and half a second after the shutter opens turn-on the blue or autofocus LED and use that (with some electronics) to fire the external flash.

You could test this in a dark room by using the autofocus led and pointing camera at its reflection in a mirror.

You cannot do it at 'normal' shutter speeds, that is why you need an external device like the gentStereo.


David
Logged
*

Offline reyalp

  • ******
  • 14128
Re: deep-sea autonomous shooting
« Reply #5 on: 08 / November / 2009, 18:15:03 »
Quote from: Microfunguy on 08 / November / 2009, 17:51:15
Quote from: glowgoose on 08 / November / 2009, 17:19:23
2. can it be programmed to trigger an external flash without flashing its own?

I am tempted to say no  ....
The manual suggests that you can control an external flash in the hotshoe independently, but it's not completely clear if this includes turning the internal one off. I would expect it too, but don't know for sure.

If flash is the main/only source of illumination, then using a long exposure and triggering the flash with an LED should not be a problem.
Logged
Don't forget what the H stands for.
*

Offline fbonomi

  • ****
  • 469
  • A570IS SD1100/Ixus80
    • Francesco Bonomi
Re: deep-sea autonomous shooting
« Reply #6 on: 09 / November / 2009, 10:33:21 »
Quote from: glowgoose on 08 / November / 2009, 17:19:23
2. can it be programmed to trigger an external flash without flashing its own?

You can always cover it up :-)

Quote from: glowgoose on 08 / November / 2009, 17:19:23
3. how long will it live/shoot on its own battery? (sorry, this one is definitely not a CHDK question, but perhaps you just know this)

It depends VERY much on the shooting conditions.

Of cource, the flash drains a lot of power.

The display is another major factor. It can be turned off but (at least in several other cameras) turning the display off disables manual focus. A possible partial workaround is to insert a jack in the A/V out plug. The camera thinks to be connected to a monitor and will turn the display off. The actual display electronics stay on, anyway, so it's only a partial workaround.

Long exposures tend to drain more power, of cource, as some electronics stay on for longer, but that's not that important.

And more obvious factors: shooting more photos will shorten battery life.

Without having a G9 and doing actual on-the-field tests, it's difficoult to say.

Other factors are VERY important, such as temperature. i have taken part in a couple of high altitude ballon launches, and in the first one (in january) I am fairly certain the extreme cold shortened the battery life. In another occasion (a launch in june) the battery lasted for wee over 3 hours, until re-entry. It was a different camera, though, so your mileage will vary.

Anyway, I don't think an external supply would be a problem. In balloon applications, where the TOTAL payload is about 1Kg, heavy batteries are a problem, but in your case I guess it's less a dramatic issue.

Anyway, an external flash will need its power supply too. Depending on the external circuitry (and on the batteries that are on-board in your vessel), you could use Canon's 12V adaptor or any other adaptor that gives the right voltage.

Anyway, if you are thinking ot making a time-lapse of a protist's motion, you might as well consider NOT using a flash, and use long exposures (of several seconds). I don't know the absolute ambient luminosity down there, so I don't know if it's really feasible, some calculations would be needed...

Unless you want a time-lapse that spans through several days, in which case I guess that shooting in deep sea AND during the night really requires a flash.

Have you thought about a LED lighting system? LEDs are usually moe efficient, so it might be easier to achieve long battery life. Of cource, the light is less strong, but again with long enough exposure you could get there.

Really fascinating project....
« Last Edit: 09 / November / 2009, 10:34:54 by fbonomi »
Logged
*

Offline Hacki

  • ****
  • 359
  • EOS 80D
Re: deep-sea autonomous shooting
« Reply #7 on: 10 / November / 2009, 02:49:31 »
Quote
at ocean depths of 2 km and deeper.

Quote
I don't know the absolute ambient luminosity down there, so I don't know if it's really feasible,

I think its safe to say that the ocean at that depth is one of the darkest places on earth. I'd be surprised if there was any light at all.

Quote
Have you thought about a LED lighting system? LEDs are usually moe efficient, so it might be easier to achieve long battery life. Of cource, the light is less strong, but again with long enough exposure you could get there.

Interesting approach. You could also use infrared LEDs, to not bother any of the deepsea-creatures.. It would only produce monochromatic pictures, but i assume most of the animals living down there arent much to look at anyway, colorwise? :)

Problem is, due to the infrared filter you'd need quite a lot, and quite powerful infrared LED's to accomplish ambiance illumination. So, some high-power white LEDs might to the trick as well, as already stated you could trigger them with a photoresistor and one of the cameras own LEDs.

Anyway. Fascinating project.

Regards
Logged
*

Offline fbonomi

  • ****
  • 469
  • A570IS SD1100/Ixus80
    • Francesco Bonomi
Re: deep-sea autonomous shooting
« Reply #8 on: 10 / November / 2009, 04:34:51 »
Logged
*

Offline glowgoose

  • *
  • 4
Re: deep-sea autonomous shooting
« Reply #9 on: 12 / November / 2009, 01:02:51 »
I am truly fascinated with the progress of this topic!
To clarify a couple basic things:
the sea at 2 km is indeed pitch black, no light except the one you bring along (discounting luminescence of them there creatures! which is so weak and rare (unless you harass them really badly, which is my part time job) we can safely disregard it for the practical purposes).
The water is very opaque to infrared, it travels through it by maybe a few feet. The visible far-red works somewhat. Actually, using such a red light to shoot "unobtrusive" videos in the deep sea is an existing project, quit exciting, too bad it is not mine.
Microcontroller is what I would like to avoid with the help of CHDK. Clearly, microcontroller solves many problems, but it is expensive to develop and an additional point of unreliability.
Current status: I talked to an engineer who develops underwater equipment, and, lighting-wise, he had a somewhat dumb, but well-working pre-made solution: a bank of white LEDs that stay permanently on. It lasts about 100 hours with the supplied battery, well over my time limits (I need just a couple of days). He says we can plug the camera into the same battery, so consider the problem of power supply solved.
He even has a pre-made housing for the camera, into which, unfortunately, my G9 does not fit by a couple of millimeters (diagonal size of the box). In this regard, the question: can you recommend another fully CHDKable camera that would be reasonably good, and fit in there?
Logged
Pages: [1] 2   Go Up
« previous next »
 

Related Topics


SimplePortal © 2008-2014, SimplePortal