The View Up Here

Random scribblings about kites, photography, machining, and anything else

Archive for April, 2010

A KAP Rig for Panoramas

Posted by Tom Benedict on 30/04/2010

I finished the modifications to my KAP rig to give it an additional axis of rotation: plan.  With the flip of a switch, I can rotate my camera 90 degrees for doing verticals or for making vertical panoramas.  The additional axis was almost entirely constructed using parts from Brooxes.  The only custom part could be replaced with a slightly modified commercial part.  Here’s the axis as-built:

HoVer Axis

The modification didn’t really add that much weight to the rig, but it did add a considerable amount of bulk. I didn’t really pick up on how much until I rotated the camera to vertical:

Subtle as a Dump Truck

I modified my transmitter by removing one of the unused axes and installing a TPDT switch in its place. Two 10-turn 5k potentiometers wired on either side of the switch give me two independent set points I can switch between. One was set to position the camera horizontally, and the other was set to position the camera vertically. The 10-turn potentiometers allow plenty of room for fine-tuning and dialing the HoVer axis in really accurately.

Sit 'n Spin

I haven’t had an opportunity to fly it yet, but that should come over the weekend.  The only other modification I still need to make is to add provisions for my safety lanyard, which is shown hanging in the wind here, and to add Velcro safety straps for the camera itself.  My previous method for strapping the camera down would interfere with the new HoVer axis.

The parts used for the HoVer axis are a Brooxes Utility Frame, a Brooxes Deluxe Gear Guide, and a custom hub.  The reason for the custom hub was to provide a positive connection between the camera bracket and the HoVer axle.  Using a plain-shaft hub runs the risk of having the set screw come loose and the camera falling free of the frame.  So I made a hub with a threaded hole in the middle.  This screws onto the HoVer axle, and a set screw then locks it in place.  If you’d like to make this modification to your rig and don’t have a machine shop to make parts in, a good alternative would be to pick up a 1/8″ bore gear hub from Servo City, enlarge the bore, and tap it #8-32 to go on the Deluxe Gear Guide axle.

One unanticipated benefit of this is that attaching the camera to the rig is very straightforward now.  The camera can rotate in all three axes, so it’s a simple matter to flip the camera on its side, roll it back a little, and really get some easy access to the tripod screw.

I can’t wait to get this rig off the ground and see what Autopano Pro does with the resulting photo sets.

– Tom

Posted in Engineering, Kite Aerial Photography, Machining, Photography | Tagged: , , , , , , | Leave a Comment »

Involved and Committed

Posted by Tom Benedict on 28/04/2010

Back when I worked for IBM, we had a second-level manager who liked to walk around with his hands on his hips and give orders that started with “I want you to commit to me that…”  It was during this time that I developed my philosophy regarding involvement in a project and commitment to a project:  In a ham and eggs breakfast, the chicken is involved, but the pig is committed.  Needless to say I rarely told him I would commit to something, considering his own level of buy-in was typically as a chicken.  I no longer work at IBM.

But the analogy holds true, and is applicable in most aspects of my life.  Take the KAP rig modifications I’m doing.  I was able to back out a number of changes I’d made to my rig over the years.  I had involved my KAP rig in those changes, but I hadn’t committed it to them.  Even when building the HoVer parts for my rig, there was a clear point when I would switch from being involved to being committed to this modification.  That moment came last night.

Toward dinner I finished making the hub that goes at the end of the HoVer axle.  0.875″ diameter, 0.300″ thick 6061 aluminum with four #4-40 threaded holes set on a square pattern 0.440″ on a side.  (I guess I have a thing for fours.)  The hub threads on and locks with a set screw, so there’s little chance of it just slipping off, even if the set screw comes loose.  Involved.

I drilled the matching pattern in my existing tilt bracket, counter-sunk the holes, and found screws to match.  Still just involved.  It’s when I sawed off the ears that attach the tilt bracket to the tilt servo that I really committed to this modification.

Leap of Faith

Ok, ok, to be fair I can still go back and order a replacement tilt frame and revert the rig to its earlier configuration.  But for what I have on hand there’s no going back.  The ears are chopped off, the bracket was milled flush and filed clean, and from this moment forth my KAP rig includes a HoVer axis.  >gulp!<

But along with the hammering heartbeat came a certain sense of satisfaction.  The modification worked.  The bracket had plenty of clearance on all sides, and by tonight I should have my rig back together and ready to roll.  When I head out to do photography this coming weekend, I’ll be able to shoot single-stripe panoramas with the camera held vertically, and in the same flight switch the camera to horizontal to do single shot exposures.  It’s a huge step forward.  And it came together as smoothly as I could’ve asked.

Now I’m committed.

Tom

Posted in Engineering, Kite Aerial Photography, Uncategorized | Leave a Comment »

More Fun with Panoramas

Posted by Tom Benedict on 27/04/2010

I got out a couple of times over the weekend to shoot more panoramas.  One that I did over Puu Huluhulu on Saddle Road turned out quite well.

Kipuka Puu Huluhulu and Mauna LoaThis was done just after 7am, with a really steady breeze coming in from the Hilo side.  The panorama consists of 21 photos taken in two stripes, something only possible with steady wind and a good solid kite.  It’s something of a milestone for me for a number of reasons:

  • I’ve been trying to photograph Puu Huluhulu under these conditions for over a year, and for over a year I’ve been thwarted.
  • This was the first time I’ve been able to really test Autopano Pro since I picked up the license last week.  I’m overjoyed by how well it worked.
  • This is the last time I ever want to depend on the wind to be just right to let me make a two-stripe panorama.

That last bit probably takes some explanation.

I long ago quit making kite aerial panoramas the traditional way, in favor of a method developed and proven by a French KAPer who goes by the name of Vertigo in the KAP forums.  His method, which he calls “Burst autoKAP” involves spinning the rig at a fixed rate and holding the shutter button down so the camera can take pictures as fast as it can store them.  With a good high shutter speed, this method works flawlessly.  Using the traditional method, a 360 degree panorama took roughly a minute.  That’s a whole minute during which the wind can change, the camera can drift, and things can get out of alignment.  Using Vertigo’s burst method, a 360 degree panorama takes my rig and camera roughly ten seconds.  The resulting panoramas stitch much more easily, with fewer errors and fewer panoramas lost to camera movement.

But your panorama can only be as tall as the vertical axis of your field of view.  Unless, of course, you shoot two stripes of photos at two different camera tilts.  A two-stripe 360 degree panorama takes my camera and rig twice as long as a single stripe, or twenty seconds.  While far better than the traditional approach (which would take upwards of two minutes), it’s still longer than the ten seconds a single stripe panorama can be done in.  Twice as long, in fact.  With twice the chance that the wind will shift, things will move, and the panorama will be spoiled.

Later in the weekend I made an attempt at a panorama of Ali`i Drive in Kailua-Kona.  The light was wrong, and the wind was gusty and light, but it let me test the concept and scout a better location to fly from when the conditions warranted it.  But the wind was too variable to allow me to try a panorama in two stripes.  Had I been relying on that technique for capturing that image under perfect conditions, the day would have been a dead loss.

So let’s go back up to that statement I made about the panorama only being as tall as your camera’s vertical field of view.  Panorama photographers have known this since they started assembling panoramas off of photographic print.  The solution is simple: rotate your camera 90 degrees so your wide axis is vertical, and you get more coverage.  It results in a slightly slower burst pattern in order to maintain the same percentage of overlap between images, but it’s far better than the 2:1 time penalty you take making a two-stripe panorama.  So the next question is how to rotate the camera in the KAP rig.

There are two approaches.  One is to change out the tilt bracket for one that holds the camera vertically.  The other is to add in another axis of motion to the KAP rig that allows the operator to rotate the camera’s orientation on the fly.  I went with this second option.

There is no pre-built kit of parts for doing this.  Most people who build a rig with this horizontal/vertical orientation axis build them on their own.  There are a handful of examples out there, but no one yet makes a kit where you can just click “HoVer Option” and get the appropriate parts.  Even so, the parts that are available from Brooxes can get you very very close.  That’s where I started.

My approach is to use my existing tilt bracket for the camera mount.  I added a utility frame and a Brooxes Better Gear Guide with 1:1 gearing in it for the HoVer axis.  This then replaces the tilt frame in my existing KAP rig, and the whole thing is re-balanced.  That was the idea, anyway.  Unfortunately life rarely works out the way you plan.

For starters, the 1:1 gearing won’t bolt straight in.  There is a clearance issue between the bracket and the side of the servo.  Likewise with everything in the normal orientation, there’s interference between the side of the servo and the utility frame as well.  Milling the required clearance was a quick job on the mill, though an equally good job could be done using a hand nibbler and a file.  I don’t own a nibbler, and the one we keep at work was… well… at work.  It was quicker to mill the parts out.

Next, there is no standard part to go from a Brooxes Better Gear Guide to the chopped-up remainder of a tilt frame.  The BBGG includes a nice aluminum gear hub that gave me the idea for how I wanted to mount the remains of my tilt frame: a 0.875″ diameter piece of aluminum, threaded to take the BBGG shaft with a set screw to lock it in place, and a bolt hole pattern to attach it to the tilt frame.  This can be made quite thin, and should keep the camera from overhanging the end of the BBGG too much.

From this point on it’s a matter of clearance.  I had to un-do a number of modifications I’d made to my KAP rig in the sake of keeping things compact.  The clearance required for the HoVer axis is quite large.  But after careful measurement, I think I can squeeze everything in.

Sorry, no pictures yet.  It’s still a work in progress.  But things are progressing well.  The rig should be finished before this coming weekend, and I hope to get a chance to put it through its paces in the air before Saturday morning.  If things work out well, I hope to get out early again this coming Saturday and catch that morning light on a new subject to go with the panorama I made of Puu Huluhulu.  But this time it’ll be using a single stripe of images and a camera rotated to vertical at the flip of a switch.

Tom

Posted in Engineering, Kite Aerial Photography, Machining, Photosynth | Leave a Comment »

Panoramic Photography

Posted by Tom Benedict on 22/04/2010

The quest for the commercially viable photograph continues.  In the previous article I looked at the photography hardware, namely the lens, camera, KAP rig, and kites I would need to make photographs that I could use for stock purposes.  Given the hardware I have on-hand, the only available method is to do panoramic photography.  This is fine by me since I already do aerial panoramas, and thoroughly enjoy the freedom the aerial vantage point offers the panoramic photographer.  But it involves an additional step that traditional single-shot photography does not: stitching software.

For the past several years I’ve happily used Microsoft Image Composite Editor for all of my stitching needs.  It’s a capable package, though a little sparse on geometric controls.  It’s also right in my price range since it’s free software.  Up until now that’s all I’ve  needed to know.  But in looking at moving from hobbyist to semi-professional, I’ve had to re-examine every aspect of what I’m doing to make sure I’m on the up-and-up.  Camera?  Check.  Kites?  Check.  KAP rig?  Check.  FAA?  Since I’m doing this from a kite and follow all the pertinent FAA regs, that’s a check too.  But what about software?

I do the bulk of my day-to-day image processing in Photoshop 7.  It’s not the latest and greatest, but it’s got two things going for it:  It’s still a rock-solid piece of software, and I have a commercial license for it.  So I can get everything done that I need to do, and there are no restrictions on what I can do with the resulting images.  But the same isn’t true of ICE.  The license agreement you sign off on when installing Microsoft ICE quite clearly stipulates that it is for non-commercial research purposes only.  Period.  So every single panorama I’ve ever made with it is not available for commercial use.

This set me on yet another quest: the quest to find commercially licensed panorama stitching software that would do everything I want.  Sounds simple, right?

Wrong.

The hunt started with a web search.  There are a lot of packages out there for stitching panoramas.  A lot.  And for every package, there are at least a handful of people who strongly like or strongly dislike it.  Some articles dismissed entire families of packages with a single sweep of their textual arm.  One condemned any package that ran on a Windows computer because no graphic arts professional would be caught dead using anything but a Mac.  I found that particularly funny, and ignored the rest of the article.  There were a handful that kept coming up, though.  So I tried most, if not all of them.

This is where things started to fall apart.  One I tried worked great, and was designed for making prints rather than immersive VR representations for web use.  But it could only handle a single stripe of images.  I do multi-stripe panoramas, so I had to drop it even though it was a promising package.  Another that came highly recommended from a number of sources simply wouldn’t run on my computer.  I use an AMD 64-bit computer running Vista64.  Seems simple enough, but that caused a good number of packages no end of problems.  This was a source of frustration since my computer and the OS running on it are now several years old.  You’d think the issues would be known by now.  Go figure.

In the end I went back to one I tried some years ago: Autopano Pro.  When I installed the demo, it would assemble panoramas, but wouldn’t let me edit them.  I almost gave up in disgust until a fellow photographer whose opinions I trust completely told me to really give it a try.  So I hit the support forums, found a similar issue, and tried the fix.  It worked.  Once the problem was fixed I was able to create, modify, tweak, bonk, dwing, and save my panoramas.  The quality was as good as, if not better than ICE.  So I bought a license.

Since then I’ve had more of an opportunity to put Autopano Pro through its paces.  It is considerably more flexible and more capable than ICE.  I also signed a commercial license agreement, so there are no restrictions on what I do with the resulting images.  I’m back in business, so to speak.

Rather than take the stance of so many of the articles I read when I started this search, I won’t demean any of the other software out there.  I won’t even mention them by name.  And I won’t make any statements about what the “best” software is.  I honestly don’t know.  But I did find one that will run on the hardware I have, will stitch multi-stripe panoramas, handles ghosting remarkably well, and offers a commercial license for someone like me who is trying to move toward being a semi-professional.  If you find yourself in the same boat, give it a try.

Tom

Posted in Kite Aerial Photography, Photography, Software | Leave a Comment »