New Product: clickPAN-SDM

edited March 2010 in KAP Gear Sources
Click for a larger image...It time to look again at how to autoKAP!
  1. Have you ever wanted to have complete control of the camera position without being limited to preset positions?
  2. Have you ever wanted to get rid of the pictures the camera takes of the uninteresting side as the camera rotates?
  3. Have you ever wanted a rig that can do RC control as well as autoKAP, and change mode in flight?

Let me tell you about clickPAN-SDM, a very successful collaboration with David Sykes (SDM) and Dave Mitchell (Scripting).

Click for a larger image...clickPAN-SDM can control 2 servo motors FROM A SCRIPT on a Canon camera running SDM 1.83 and above.

The script sends commands to the servos via the AF assist lamp (which you don't need when shooting outside). clickPAN-SDM has a sensor which is velcroed over the camera AF assist lamp to receive the script instructions, as well as a battery input and 2 servo outputs. The unit can interface to standard servos (can move them up to 180degrees if the servo can support it) as well as servos modified for continuous rotation.

There is complete flexibility, so if you want to design a rig that uses 2 standard servos and only "scans" 90degrees of the sky then there is no reason why not. The power is in the scripting!

This is a lot to get your head round! There is more information here:

SDM Download 

clickPAN-SDM Device 

clickPAN-SDM Scripting 

At the moment the documentation is aimed at users who are already familiar with installing SDM and running scripts. The material doesn't cater as well for those who are new to SDM or who want to use this device for other applications or take it further.

Earlier I mentioned you could have a RC and autoKAP rig, the commands are all there and tested but no-one has built the rig yet! Its only once people start using clickPAN-SDM that their ideas will show its full potential!



  • hmmmmmmm . . .

    I may be sold already.
    I have no time
  • James, David and Dave,

    Very interesting.

    Getting close to fulling part of my "Auto KAP" requirements posted here.

    Brief skim of the materials and links looks worth exploring....few initial thoughts....

    - Could develop a start kit of KAP scripts that cover common KAP scenarios (directional targets, quick 360 survey, ....)
    - I assume from a quick read ..... in the field the KAPer would select a specific script given the intended KAP session .... OR
    - A single multipurpose script with selectable parameters
    - Love to find a way to tie in the GPS data logger out put (simmilar to Cris's Q Starz-1300 GPS) and feed script parameters to basically "move the servos to keep the camera pointing at the target +/- 30 degrees...
    - Like the little clickPAN-SDM device. Neat and efficient
    - Question - using the auto focus light in the camera to "read" the SDM control outputs...does it impact the focus....or do you need to use the SDM command to fix focus at infinity?
    - Weight considerations....looks does it compare with the autoKAP AuRiCo (like for like)?

    Bottom line - like where this is going...may have to shell out a few clams to give it a try.

  • Interesting. I use standard CHDK, not SDM, but might give it a go. This sounds like it could have some use in PAP. I know there is already a Gentles PAP toy that allows control of pan, tilt and shutter with a garage door opener. Could this new toy be used for AutoPAP?

  • edited March 2010
    Could develop a start kit of KAP scripts that cover common KAP scenarios (directional targets, quick 360 survey, ...) I assume from a quick read ..... in the field the KAPer would select a specific script given the intended KAP session ... A single multipurpose script with selectable parameters
    Dave's scripts already provided use parameters to help have a generaic script for a rig type, but you are right, once you decide what kind of servos you want you are free to script and play. The standard configuration is a stanadard 90degree tilt and a continuous PAN but there is nothing to stop you using 2x 180degree servos or any other combination.

    Love to find a way to tie in the GPS data logger out put (simmilar to Cris's Q Starz-1300 GPS) and feed script parameters to basically "move the servos to keep the camera pointing at the target +/30 degrees...
    Dave Sykes use of the AF lamp gets data out of the camera, getting data in is still through the USB, and currently limited to the simple switching achieved with gentLED-CHDK2

    using the auto focus light in the camera to "read" the SDM control outputs...does it impact the focus....
    The AF lamp is only used by the camera in low light conditions.

    Weight considerations....looks light...
    Around 7grams, half the weight is in the cables to the sensor and battery.
  • As well as sample scripts for the most popular rig-type (standard tilt servo and 360-modified pan servo), there are also samples for rigs with a standard tilt servo and a 1:4 geared pan servo (like Cris Benton's) and for rigs with two standard servos (possibly geared) with 180 degree motion (I think Dave Wheeler did a rig like this, where for some shots the camera is upside-down). If anybody has a rig not covered by the samples I'll happily write scripts for it!

    The samples are all single-purpose - that is they go on doing the same thing over and over (see the videos here). However it's pretty simple to modify them to do more elaborate things - take the occasional video, or zoomed shot for example. It's also not difficult to write a more elaborate script that starts off doing simple panning, then spends a few minutes doing pan and tilt, etc (along the lines of TUCIT's KAP Session Menu). The scripting guide offers some help in this (though it's in its early stages - I hope to make it more comprehensive).

    As for AutoPAP, I think my clickPAN-SDM rig could be converted to AutoPAP pretty easily by taking off the Picavet and replacing it with a plug or socket to fit the top of a pole.
  • I've only had a chance to quickly review the documentation but I have to say that this appears to be great work and a significant addition to KAP/PAP options. At first, I was interested in applying it to PAP, since I intend to do a great deal more pole work. But now I'm starting to consider how I might use it in KAP, too. I am pleased with my current AutoKAP rig with the AuRiCo but the flexibility of having even more script control and more precise pointing is very appealing.

    Thanks for all your effort!

  • Dave, I was thinking this type of system would be cool for some somewhat modular control. For instance waiting for a camera controller to take a set number of shots and then communicating to rig controller to move the PRT rig to move to a new position. I'm probably alone here, but does that sound easy?
  • edited March 2010
    We had hoped that the first US stock of ClickPAN-SDM would be here in time for this announcement, but they aren't. Probably Monday.

    I'll honk when I've got some.
  • When James told me about this project a year ago I was somewhat skeptical. The reason for that was that I had some headaches installing CHDK on several types of Canon cameras. Now that James send me the new clickPAN-SDM I have to revise my opinion.

    With version 1.83 of SDM installation of the new software is super simple. And the documentation written by David, Dave and James is excellent (but you must read, even the small print).

    Seeing the size of the clickPAN-SDM I decided to sweep the dust off of an abandoned single side rig. Now that I don't need a USB connector anymore I could move the camera closer to the tilt servo and shorten the camera support. Now the door that gives access to the batteries and the SD card can be opened without taking the camera of the rig., on Flickr">KAPshop's clickPAN-SDM rig

    So far I'm very happy with the clickPAN-SDM. Lot's of things to play with. Thank you, David, Dave and James.

    One little thing: can David make a 1.83KAP version? That is one with Dave's clickPAN-SDM scripts instead of the scripts that come with the standard version. That would make things even better.
  • TGRAN,
    although the question is directed to Dave, can you explain a bit more what you mean ?
    Quite honestly, it is likely you can do anything that you practically need !

    No need for 'instead', I will just add them to the Scripts folder in (if Dave agrees).
  • David, I'll read up more but I'm loving the possibilities, I agree there are a lot of different directions to go. Very exciting!

  • edited March 2010
    I have put the clickPAN SDM on the top of my list of KAP things to buy, wide angle lens and tube is now 2nd.

    Just thinking out loud here. How hard would it be to have one with 3 servo outputs? The 3rd output could be for a HoVer servo. That way you could do a KAP/PAP session with the camera horizontal or vertical. If the script would support it, change the camera from horizontal to vertical and back during the same KAP/PAP session. I do not know of an auto controller that does it, with out making one and that is over my head. May not be a large demand for it. For me it would be the auto controller to have. There is a topic as to what is the best pano format, why not both?

  • poolking,
    3 outputs? Nothing is impossible, I'll add this to my list of "whats next", but nothing will happen until people get some experience with the current offering which will point the way to what people want in a future product.
  • Dave, few more thoughts, questions on the click-PAN-SDM.

    In the manual (which I have only skimmed) it looks like you have a calibration process to test out individual rigs.

    I understand the basic command and general approach.

    Has anyone estimated the battery drain impact with using the focus assists light to pass controls to the servos?

    Few questions on starting up in the field for KAPing with the click-PAN-SDM.

    I am assuming the following workflow in the field:

    - Upon starting up the camera and selecting the SDM script the camera slews the rig to a standard "standard" home position (please confirm).

    - The rig could be manually synced up to a North compass heading. (please confirm)

    - The KAPer could select from a list of predefined scripts for the desired target. Something like the following example:
    - Shoot East (90 degrees) + - 30 degrees for 10 minutes with shot every 5 seconds
    - Do one 360 pan with tilt pattern with shot every 5 seconds
    - Shoot East (90 degrees) + - 30 degrees for 10 minutes with shot every 5 seconds
    - repeat (till battery dies or card fills).

    Key to this work flow working is the assumption of a) camera slews rig to a standard home position and b) KAPer can align the home position to a North (or other predefined) compass heading.

    Please confirm (or correct) my thinking and or propose an alternate "in the field workflow".

  • edited March 2010

    I can't answer the battery drain question at the moment, but I'll run some tests.

    The answers to your workflow question depend on what sort of rig you have. With a 360-modified servo it's likely that you can adjust clockwise and anticlockwise pans so they are close to equal (ie having set the rig so the home position is north, you can do a series of pans and come back to near where you started). However, over time there's bound to be some drift so on long sessions things might gradually get of of alignment. There is no 'home' pan position on such a rig, merely the position the rig is in when you start the session.

    With a standard pan servo and 1:4 gearing (like Cris's rigs) then absolute repeatability is pretty much guaranteed - on such a rig there is a home pan position just as there's a home tilt position. The supplied 90PT.BAS script for example can be set to pan back and forth between any two angles.

    As for syncing up to a North compass heading, that's not possible. If the wind swings round then the kite, line and rig will turn too. Any 'home' pan position is therefore relative to the direction the line is in. So if the wind is pretty constant then you could run a script that panned between North East and North West for example - by starting with the rig pointing north and setting the pans to move through +/- 45 degrees. But a sudden change in wind direction would change that.
  • I was aware of battery drain considerations when designing clickPAN-SDM. For this reason Dave Sykes put a modification into SDM to reduce the AF Lamp activity to a third compaired with the "standard" method of using the AF Assist Lamp available in builds before 1.83.
    So we tried, lets see what Dave Mitchell thinks we achieved!
  • Speaking of builds, you previously said clickPAN-SDM works for all models compatible with version 1.83. You website documentation goes to 1.82. Which models of camera can accept version 1.83? I use the old A570IS but may be moving up soon.
  • I can confirm the A570IS (software version 1.01a) works with SDM 1.83. See the rig picture above.
  • I hope beyond hope that Canon, as well as Nikon, Pentax, Ricoh, Casio, Sigma, and EVERY other camera manufacturer reads this. Really really hard. This collaboration has basically handed programmable mount control to practically every Canon Powershot camera. This is a major deal! KAP applications aside, this means you could do timelapse wildlife photography and pan 360 around a vantage point for doing head counts for population studies. You could do timelapse Gigapan style pano movies. I haven't made the switch to SDM, but if SDM retained the motion sensing code from CHDK, I'd be curious to see if you could program the mount control to track movement and plop it in the center of the frame. This has all manner of wildlife study applications as well. I know these are a tiny fraction of the market share for compact cameras, but who knows how the market will move once this kind of stuff is available?


    Ooooh the possibilities...

  • Tom,

    not only does SDM 'retain' the movement-detection code .. it is a more advanced version with additional features.

    You can display the location of every monitored point (with or without the grid overlay) and there are features that detect the amount of 'stillness' (camera shake, for example) or no motions for a defined time.

    Maybe that will work with KAP, maybe it will not.

    Tracking an object and creating Gigapan images also occurred to me .. but I can only do so much :-)

  • Muahahahahahahaaaa!

    Using that for KAP, it's possible. I've wondered how useful it would be to have a script that, when it saw a +5V at the USB port, would wait until things were still before taking the picture. Rather than doing active image stabilization, it would do passive stabilization.

    Christian Buil, an astronomer in France, took a similar approach to his poor man's adaptive optics setup. Rather than use an AO system with a tip-tilt mirror and a deformable mirror, he used a straight-through camera that would take many many images very very quickly. After the fact, the blurry frames are tossed and the sharp frames are co-added. The resulting image, of course, has a great deal of dynamic range, and is as sharp as your rejection process. Of course this doesn't work on dim galaxies or distant stars, but for planetary astronomy it works like a charm.

    But it sounds like it might be possible to let SDM do the rejection on the fly, and only trip the shutter if things aren't jiggling around. Co-adding I'll leave for another day.

    Ok, gotta play with this.

  • edited March 2010
    have you seen this? It's not part an official part of CHDK or SDM but I think it's a pointer to the future - being able to do this sort of real-time video analysis in a cheap point-and-shoot is pretty amazing.
  • MAN this learning curve is starting to look steep again...

  • Dave, thanks for the feedback on my post above.

    Per comment "SDM script the camera slews the rig to a standard "standard" home position...." and "The rig could be manually synced up to a North compass heading." What I was thinking was for the KAPer to manually line up the rig to face north (manual compass), then select appropriate KAP script for a given target. I understand the PAN servo can not be "homed" but with give starting point (oriented North) this should not be an issue. Per the comment of wind shifts.....agree 100 % (especially on long high flights)....thus this is why I would define an error range on the pan angle (+- 30-60 %).

    Future requirements
    - have the camera script read inputs from a GPS position and heading and then send commands to the servos to adjust pan (maybe tilt) to keep the camera on the KAP target....even track to multiple targets from a relative current GPS position and heading (facing). With this could pre program a KAP script to track several targets even while you are walking around even with the wind shifting.

  • Regarding the real-time video analysis, you have to realise that after booting, the camera continuously executes the image-processing code.

    It does not do anything else.

    None of the normal shooting features or menus are coded.

    It works at about five frames/second but as there is no way of making aperture and shutter-speed constant (required by the algorithm) is of limited practical use.

    Olivier simply wanted to show how fast his algorithm is.

    Nevertheless, it is interesting :-)

  • I'll bet that's OUR Olivier -- Humez. Amazing as usual, if so, as you will see at KAPiNED.
  • Battery Consumption Report for clickPAN-SDM on A650

    I did a ground-based test today. My A650 ran the 360PT.BAS script for 3 hours and took just over 500 photos (I had set the delay between shots quite long) before I stopped it. I think it would have gone on for another hour at least. Note that the A650 takes 4 AA batteries - I was using fully charged Eneloops (though they are several years old and have been recharged many times).

    I'll repeat the test on my A570 this afternoon.
  • Battery Consumption Report for clickPAN-SDM on A570

    I did a second ground-based test today. My A570 ran the 360PT.BAS script for 2 hours and took just over 300 photos before it stopped because the card was full. At that point the batteries were pretty exhausted. Note that the A570 takes 2 AA batteries - I was using old Eneloops again - this time they may not have been fully charged when I started.

    So it looks as though the clickPAN-SDM will do a reasonable AutoKAP job on both cameras.
  • Broox,

    no, he is Olivier Barnich .. a researcher in computer vision.

Sign In or Register to comment.

Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

In this Discussion