SVP2014 – the wind-up

Today is Sunday, November 2, 2014, which means that the 74th Annual Meeting of the venerable Society of Vertebrate Paleontology will start in less than 72 hours. Hashtag #SVP2014, by the way. Depending on how you count the public lecture on Tuesday evening, things start day after tomorrow in the evening or one day later – at 8 a.m.!

As is usually the case, the last minute preparations are crazy, but this year the crazy is extreme. SVP comes to Berlin, and I am on the Museum für Naturkunde’s Host Committee. That’s added stress to the usual, with the latter being two posters (one as first author, one as second who has to do most of the work due to our division of labour on the project). The first is finished, the second still in the making. Ugh! But both posters deal with cool stuff, and I am very much looking forward to my colleagues’ reactions. Obviously, I’ll post about both here, too.

What else will be up this week? For one thing, EVA Berlin 2014! Where I will give a talk on Wednesday morning (yes, ideal timing *sigh*) on how a lack of standards for documenting how you 3D-sscan something means that nobody documents anything, which invalidates many 3D-models for research. Ugly topic, and I hope my talk will be a sort of wake-up call for many players in the field.

And then, obviously, there will be a bit of free time to talk to all those many friends and colleagues I only see sporadically. And there will be beer, and dinosaur talk, and anecdotes about crazy professors, crazy students, crazy excavations – i.e., there will be vertpalaeo-fun :D

So, back to poster making…. here’s a sneak preview.

tail recon

Photogrammetrically derived 3D model of the tail of Citipati osmolskae with partial muscle reconstruction.

Posted in 3D modeling, Berlin, Citipati, classic CAD, Conferences, Digitizing, Dinosaur models, Dinosauria, Maniraptora, Oviraptorosauridae, photogrammetry, SVP 2014, Theropoda | 3 Comments

Photogrammetry tutorial 7: multi-chunk project handling

In palaeontology (and many other disciplines) you often deal with specimens that you wish to capture in what I term “720°” – all around (360° around it for the vertical axis), and bottom and top, too (360° around the transverse axis). Because physical objects tend not to float stably on their own in thin air, this usually means taking one or more sets of photos, flipping the specimen over, and taking more photos. As a consequence, you will capture a lot of background in your photos that moves in relation to your specimen between the photo sets, and this background can then cause trouble when you build the model.

As I keep pointing out, e.g. in that paper on photogrammetry I recently published with Oliver Wings, it’s all no problem if your background is featureless. The software will not find points and ignore it, so you can simply toss all your photos into one chunk and be done with it. Neat trick – but it needs to work! (bonus points if you recognize the quote)

But what if a featureless background is not possible? Let’s say you’re dealing with a sauropod femur! It is highly unlikely that you can just lug it around and position it at your will in front of a greenscreen. Normally, you’ll be lucky to be able to lean it against a wall in various positions (long bones should be kept upright if possible and if sturdy enough, because that minimizes the bending moments on the shaft). Or it will have to be laid flat on the ground or a table, with some styrofoam or sandbag supports. In either case, there will be lots of stuff very close to the bone that you can’t turn into something the photogrammetry software doesn’t find points on. Like this:


The proximal end of MB.R.2694 (field number ST 291), a Giraffatitan brancai femur from Tendaguru, Tanzania, today to be found as a kind of stumbling block in the MfN’s famous bone cellar. This little chunk weighs (guestimate) 40 kg, so it really is not something I want to toss around a lot. Additionally, as hard as the fossil is, its own weight is sufficient to cause local damage when you put it on something hard, like the floor. So, the black mats you see in the photo are a must, and they are structured and dirty enough to give plenty of features for the software. Plus, they are necessarily so close to the bone that whatever I do, they will be in focus on the images. Tough luck.

So, what to do? I took two sets of photographs, one from each side, making sure that I get good images from the sides. That’s a thing to keep in mind: do NOT take a lot of photos from above, but rather lots from the sides. This gives you a lot of overlap near the margins of each model, so merging them into a complete model is easier.

To create a complete model I could now mask all the background in all the photos, toss them all into one chunk, and have the software align them all. Good, but not perfect, because masking is a lot of tedious manual work. With a smooth, uniform background it would be much easier, because the magic wand tool would make masking easy, but that’s not to be in this case.

Thus, I had Photoscan calculate the alignment and dense cloud for each set of pictures, which is a piece of cake. A key point: I did so using two chunks in one overall file. Each dense cloud got trimmed down to the bone, the equivalent of the masking I could have done on the images, but a lot faster! Basically, I rotated the models so that I was looking at the points representing the ground in perfectly lateral view (so they form a line on the display, not an area), then used the box selection to cut them away. Additionally, I cropped the rather rough margins of the bone model. All in all this takes about a minute or two per model, compared to some 15 or more minutes for masking the photos. And the more photos there are, the bigger the difference.

So, there I was with two half-shells of the bone, and in need of a neat way of matching and merging. And so the fiddling started: In both chunks, I went looking for three easy to find spots on the bone photos, and placed markers on them. In doing so, I made sure that the same spot on the bone got identically named markers in each chunk. That’s easily achieved by creating the markers in the same order, so they get named “point 1″, “point 2″, and “point 3″ automatically. It can help to place them in one chunk, then take a screenshot of a photo on which they can be easily seen (or of the dense cloud), and paste it in a graphics program. That way, you can go back and forth between the photos of the other chunk where you need to find the same spots again and an image of where they are without having to search around in Photoscan.


Here’s an example of such a screenshot, you can see the green marker I put on a dark colored spot that I believe is easy to find on other photos.

In the end, the first half of the model, the first chunk, looked like this as a dense point cloud:


You can see four markers for alignment here. What’s with all these “target” markers? That’s for another post…. For now, it is enough to see that I spread the makers around the circumference of the bone, which means that aligning by them will not leave any end of the bone far away from a marker and thus susceptible to errors caused by tiny inaccuracies far away that simply add up. Misalignments typically mean that things are off by a tiny angle, that the two triangles formed by the two sets of three markers do not match perfectly – and distance from the marker points thus means that the tiny angle inaccuracy adds up to a lot of local separation. So it is a good idea to use more than three points, and to place them at the corners of your specimen.

Next up, I let Photoscan align the chunks based on the markers. If I am lucky, all comes out perfectly. Then, I turn “Show aligned chunks” on, so that I can see what the two dense point clouds together look like. This tells me not only about the quality of the alignment, but also if I cut away enough of the rims of the models.


Here, you can clearly see how well my markers matched (by the fact that you can’t see that there are two of each marker), and how the two models match nearly perfectly. There are a bunch of erroneous points that needed trimming (easier done in the individual chunks BEFORE merging), but I was rather lazy – Photoscan Pro is very good at ignoring those little floating islands you see.

And here’s now the mesh in all its glory! 28.25 million polygons :)


Obviously, I photogrammetrized the other parts of the bone, too, and will digitally repair the entire bone. But that’s for another post.

Posted in 3D modeling, Digitizing, Dinopics, Dinosauria, Giraffatitan, How to, MfN Berlin, photogrammetry, Sauropoda, Sauropodomorpha | 2 Comments

A more detailed take on Pinus macro photogrammetry

Previously, I described a few photogrammetry tests I did with a macro lens, but I didn’t go into the technical details much. Some people have asked me how exactly I took the photos, so here’s the detailed description.

First of all, let me point out a few things that make photography with a macro lens different from normal lenses. There is, for one thing, not really a difference between focusing and zooming with a macro lens. As you focus, the lens extends or retracts and the angle of view decreases or increases. Therefore, you can’t just select your view, then focus – you may end up with significantly more or less on your photo than you intended.

This means that I often find it easier to just leave the camera alone, and move the specimen into focus (or move the entire camera), rather than fiddle with the focus/zoom. That’s easily achieved if the camera is on a tripod, and the specimen is placed on a support that itself slides easily on the table. At the office we have Formica tables and Formica cupboard boards, so I put my turntable on a board which slides well on the tabletop – I can pull or push it with one finger while sitting next to the camera on its tripod. Or, in the case of specimens that do not need to be handled very carefully (such as a modern pine cone), I simply move the specimen around as I please.

In the case of the fossil and extant pine cones I posted on recently, there was no need for a turntable. Nor for a sliding support. I simply manually rotated the specimen the way I wanted to.

Another difference when using a macro lens is the background. Normally, you should try for a undefined background of very even colour, which is either completely ignored by the program, or can be masked using a magic wand tool. Using a macro means that the entire background is automatically totally out of focus – even if you in fact used a Persian rug the program will not find any points. Makes life a bit easier :)

So, let’s have a step-by-step playback of what I did, using this detail model of one cone scale I made.


  1. place cone on table
  2. place camera on tripod, with 50mm macro lens
  3. extend macro lens all the way
  4. shift tripod back and forth until cone is roughly in focus (checked via LCD screen, you can also peek through the viewfinder)
  5. select f-stop (high, in this case 14, the maximum the lens will do) for a good depth of field
  6. shift (carefully!) the cone until the scale is perfectly in focus (I zoomed in on the LCD for this)
  7. use the preview button on the camera (usually hidden and unlabelled under the lens attachment on the left hand side) to see what the real depth of field is like
  8. take photo by releasing the shutter via a soft touch on the touch screen (alternatively, us a remote and ideally a mirror pre-release)
  9. turn cone a tiny bit, check focus
  10. repeat 8 & 9 until target sufficiently covered; if necessary alter tripod height for additional angle.


Posted in 3D modeling, Digitizing, photogrammetry, photography, plants | Leave a comment

Addendum to SV-POW!’s “SO close”

Mike Taylor recently sent me an email asking for a larger version of a figure I once published in a book chapter. Naturally, I promised to send it to him. But, being away from my computer, with email available only via my phone, I couldn’t send it right away. And promptly forgot.

Thus, Mike had to go ahead and publish the planned blog post without my figure. Here’s the link.

Now, I finally remembered, and through sheer persistence managed to find my original submitted files in my sorry heap of data backup. Here goes – starting with a small part of a big figure, but a part that shows nearly all that needs to be shown to make Mike’s point:


This is a CAD model of Diplodocus I built long ago. It is rough, and takes air spaces within the body into account in a fairly rough way, but – and that’s a big issue with dinosaur models – there is enough soft tissue mass modelled onto the tail, especially its base.

I used a rather complex multibody dynamics program to do all kinds of shit with the model, but first of all I had it calculate the position of the center of mass (COM). I won’t bore you with the full story; suffice to say that I put a small white dot in the figure to show the result.

Yes, the COM is that close to the hind limbs.

Now, one model alone would be bad practice, so I ran a bunch of variations. As long as my assumptions about density and volume stayed reasonable, the COM stayed really close to the hind limbs.

The next thing I did is look at another sauropod, one with a quite different overall look: Giraffatitan. Unsurprisingly, a much shorter and thinner tail and much longer and stouter forelimbs meant the COM came to rest somewhere else entirely. You’ll see in a moment…..

Then, I went ahead with the work I had set out to do originally: look at the ability of the two sauropods to rear into an upright stance. Not just rear up and come back down with a thump! right away thanks to gravity, but the ability to adopt a bipedal pose that puts the head high up for feeding on large trees. A position that must thus be held for quite a while, which in turn requires that it is inherently stable, that you can get there easily, and that you can get back down speedily, too.

Why inherently stable? Well, think of a ladder and how you behave on it. When you are high up on a ladder you will either be careful and restrict your movements, or you will fall. If you’re positioned so that the tiniest motion unbalances you, there is no way you can do useful work for a prolonged time. For a sauropod poking its head into a mass of branches, grabbing them and pulling vigorously most certainly was an activity that required it to be posed so that it did not constantly worry about toppling over.

Why easy to reach? Because if it is really hard to reach, harder to reach than a similar pose is for an elephant, then the effort to get there makes it so difficult to use that it is no good for regular behaviour. Sure, you can do the weirdest stuff, like walking on your hands. But every day?

And why is getting back down so important? That is in fact the simplest point: if a big theropod ambles by, you do not want to spent several minutes carefully letting yourself back down into a pose in which you can deal with the threat.

So, I modelled on, and here are the figures as the appear in full in the article:


This is one way Diplodocus can easily get into an upright position. By pushing its butt backwards a bit, flexing the knees a tiny bit, the COM comes to rest right above the hind feet. Now, all it takes is a very slight rotation in the hip joints, easily achieved by the strong caudofemoral muscles of both sides acting together, and the entire animal minus the legs starts rotating. When the tail hits the ground (slowly and softly) it is time to stop. There you go – feeding height doubled, or tripled, depending on how mobile you believe the neck is in extension.

Back down is also easy: just a push forward with the forelimbs to get some momentum, and slowly relaxing the caudofemoralis muscles and the animal is back down in a few seconds, ready to fight or run away (well, amble away that is).

In between, all is well balanced, because the limbs that can counter a slight imbalance easily, the hind limbs, are attached to the body at roughly the height the center of mass is at. Motions of the COM can therefore be countered quickly, before the shit hits the fan. If there was a long lever arm between the point where the animal can influence the trunk’s position relative to the limbs and the COM, it would be a very difficult task to fiddle things into equilibrium.  Additionally, in lateral view the angle between the line connecting the edge of the support area and the COM and the vertical through the COM is larger the further down the COM is, meaning higher stability as well.


Giraffatitan left, Diplodocus right. Guess who’s standing stably, and who’s more like a drunk on a ladder?


Various poses I tried for Giraffatitan. The COM is shown by a tiny sphere in the middle of the body. High, high up, far away from the hip joints. This doesn’t look good for balance. Add to that the emaciated tail base and thus weak caudofemoralis muscles of Giraffatitan, and you get an animal that was really bad at rearing up and staying there. Diplodocus, on the other hand, was (I concluded) well capable of rearing up and staying there for an extended feast.

So, what does that mean for Mike’s bipedal Diplodocus? For one thing, the COM is in the right place. The other thing is that a bipedal pose would have required somewhat flexed hind limbs. Not something I guess the animal did for a long time. And there are lots of issues with walking bipedally, not least the issue of retaining steering. Obviously, one can use various tricks for keeping the body position stable, but why bother if there is an easier way? Even a paltry 10% of body weight supported by a forelimb placed well forward of the hind feet would help a  tremendous lot with going in the actual direction you want to go. So bipedal Diplodocus yes, but not regularly!

Oh, here’s the full citation for the paper; email me for a PDF.

Mallison, H. (2011). Rearing Giants – kinetic-dynamic modeling of sauropod bipedal and tripodal poses. Pp. 237-250 in Klein, N., Remes, K., Gee, C. & Sander M. (eds): Biology of the Sauropod Dinosaurs: Understanding the life of giants. Life of the Past (series ed. Farlow, J.) Indiana University Press.



Posted in 3D modeling, Biomechanics, classic CAD, Dinosaur models, Dinosauria, Diplodocus, Giraffatitan, Sauropoda, Sauropodomorpha | 11 Comments


While I was on holiday my latest paper finally went online. It is a joint effort with Oliver Wings that details some good approaches for photogrammetric 3D modelling specimens. You can find the PDF at this link. The web version at the Journal of Palaeontological Techniques is not up yet, but should be within the next few days, starting with this link.

The paper is by far not exhaustive; there are plenty of other approaches to getting good models, and I am sure we could have added 50 more tricks for getting better models. However, the basics are all covered. Basically, if you read this paper beforehand, or if you have it handy while doing photogrammetry, you should be able to consistently produce good-quality models of specimens either in your own collection or during travel.


I have been toying with the idea of writing such a paper for quite a while, even got to the point where I started sorting my thoughts by pre-writing parts of the paper as a series of posts here (Part 1, part 2, part 3, part 4). Just then, Oliver Wings contacted me with the suggestion of writing a Photogrammetry How-to paper together. He sent along a list of dos and don’ts of his own, and we quickly set to work. Totally unusual for paper publishing: the review was the fastest part, the writing and revising was done nearly as quickly, with the figure creation barely longer. Proofing took much longer, and the wait for the paper to appear was the longest part – not because of a printing backlog, but because small journals run by volunteers are typically not too well staffed during the holiday/fieldwork season. Overall, despite the “summer hole” delay, things went very quickly.

We decided very early on that this paper would go to a no-cost open access journal, for the obvious reasons: we did not have much publishing money, and wanted the paper to be available to as many people as possible.

A really cool thing is that I was allowed to use images (and models made from them) taken during my 2013 American Museum of Natural History visit for the blog posts and paper. A huge THANK YOU to Carl Mehling and Mark Norrell for that!


Many thanks also to the two reviewers, Matthew Wedel ( and Stuart Pond (! Stu was the “expert reviewer”, the person who knows a lot about photogrammetry, whereas Matt was the “novice” – someone with little knowledge of photogrammetry, thus our target audience. The editors made an excellent choice :)

In the end, Stu suggested a re-structuring that we didn’t do because we wanted the paper to be less a scientific work, with words used as efficiently as possible, and more a how-to guide. The latter requires that some things get repeated in several sections, so that readers do not have to scroll back and forth between sections. Also, if you’re a photogrammetry novice, quite often you will not even realize that another section of the paper has some helpful info, and thus miss out on in when you consult the paper during your work.

We also received unofficial reviews from several people, for which we are very grateful. The more people read such a paper before publication, the better. It eliminates stupid errors that would later confuse lots of people.

A very special thank you goes to Emanuel Tschopp and Peter Falkingham, who handled our submission as editors. They did a very fine job! If you have a palaeo-technical paper to submit, consider JoPT or Palaeontologia Electronica’s Technical Article section. Two excellent choices :)

Oliver himself proved to be a very nice person to work with on this manuscript, as he was only interested in producing the best-possible result and dealt with the paper promptly, despite having a lot on his plate. That’s how it should be among co-authors :) I’ve published with him before, most notably a paper (Wings et al. 2007) on a tracksite in the Turfan Basin (PDF here), and I’ll happily do it again. Thank you, Oliver!

My biggest regret, now that the paper is out and immutable, is that my highly successful experiments with photogrammetry with a macro lens came too late. I would have loved adding an example.

So, check out the paper if you want some pointers on photogrammetry – or just come to the DigitalSpecimen 2014 conference! There is a special photogrammetry workshop by Brent Breithaupt and Neffra Matthews, and plenty of talks about it. :)



Posted in 3D modeling, AMNH, Conferences, DigitalSpecimen 2014, Digitizing, Dinopics, Dinosaur models, How to, Khaan, Maniraptora, Open Access publishing, Oviraptorosauridae, papers, photogrammetry, photography, Theropoda | 1 Comment

Photogrammetry with a macro lens

So far, most of my photogrammetry efforts have dealt with specimens in the several centimetre to meter range. Bones, skulls, entire skeletons. The smallest models I created were of ammonites – more on that currently on-hold project later – with diameters of several centimetres. “several” here meaning more than 4, i.e. 40 mm. That’s a range where my kit 18-135 mm lens is hard pressed at the lower end, because the specimens are so small on the image that the alignment in Photoscan often fails unless I let the program use the image background. And that causes further bother later on if I need to digitize a specimen in 720°, as I call it: 360° around horizontally, and also all around (another 360°) vertically – a specimen that I want to capture the “bottom” of, too. In contrast, a specimen embedded in rock would be only 360°, because I do not wish to digitize it from below.

If now a specimen makes up only a small part of each image, its surface isn’t well resolved in them, and Photoscan is hard pressed to find features to correlate with other images. That means that I can’t just mask the background out of all images and toss them all into one chunk for alignment, because the angle difference between many images will be too big for Photoscan to work it out based on few matching points.

Quite the obvious solution is the use of a macro lens. Which requires having a macro lens. And that’s as far as I got – until Wednesday! Here’s how I got by a macro lens and what I did with it:

Continue reading

Posted in 3D modeling, Conferences, DigitalSpecimen 2014, Digitizing, photogrammetry, photography, plants, raves | 3 Comments

How to easily de-conflict event timings for conferences

Today’s topic has nothing to do with dinosaurs, but it is something some of you may encounter every once in a while. Let’s say you’re planning a conference with many events – talk sessions, poster sessions, workshops, whatnot. Obviously, you’ll ideally time all events so that everybody can take part in all events they want to. However, it is usually not possible to have only one single session at any one time. As soon as you have dozens of people attending you’ll end up with more events than time.

So now, you’d ideally find out who wants to attend which events, then find the combination of timings that reduces collisions to a minimum. But how to do that? How do you find out which potential collisions occur least often?

I am facing this problem right now for the DigitalSpecimen 2014 conference, where up to 100 people may wish to attend up to 8 workshops each. Here’s how I solved the problem with a little very primitive EXCEL magic.

1. Polling the attendants
I asked people to register for the workshops they WANT to attend. I assume that they all answer truthfully. Thus, I know who wants to attend what combination of events, i.e. which events should not be at the same time for that person.

2. Hacking the data into EXCEL
I now made a table in EXCEL where each line is for one attendant. It has one column for each workshop. I enter a “1″ if the person registered for the workshop, and a “0″ if the person did not register. This is what things look like now:


Don’t mind the gap – I only use it to separate MfN employees from the rest. Sums and all work across gaps, too.

3. Computing collisions
I now added further column, one each for each possible conflict. Thus, I no have columns titles 1/2, 1/3, 1/4….. 2/3, 2/4, 2/5…. 3/4….. and so on. (Remember to format the cells as “text”, otherwise the program will turn these entries into dates or whatnot!)  Now, I enter formulas into the cell from the row with the first attendant: =PRODUCT(Cell1;Cell2), with Cell1 the cell that holds the info for that attendant on the first of the two possibly conflicting workshops, and Cell2 the info on the other one.
If the person registered for both workshops, the result obviously is 1*1=1. If she or he registered only for one or for none of the workshops, the result is 1*0=0 or 0*1=0 or 0*0=0. Thus, if there is a potential conflict, the result is 1, if there is none the result is 0.
Now, I know which collisions occur for the first person whose data I entered. By pulling down the bottom right corner of the field across all rows that hold names I can make EXCEL apply the same formula for all of them, and quickly have the info for all attendants.

4. Summing up collisions
To find out how often each collision occurs, I now add a row of sums at the bottom. Under each column of 0s and 1s (for no collision and collision) I enter =SUM(Cell3:Cell4), with Cell3 the top-most data cell in the column, and Cell4 the last one. And presto – I get a nice row of numbers that tell me how often each collision would occur for a given pair of workshops.
Here’s how that looks:

And now, I can see at a simply glance which events I can place at the same time without pissing off too many of my esteemed colleagues, and which pairings I better avoid! In this case, Workshops 1, 3, 5 and 7 should all be planned to not run in parallel, whereas a WS 2 – WS 8 collision would hurt only few people.


Posted in Conferences, DigitalSpecimen 2014, How to, non-palaeo | 1 Comment