3D Rock Art Scanner - Computerphile

[@lurandir8230]

In game development we usually name it voxels. Also pixels does not represent continuous data but rather points in space (as was explained in the video to be opposite). Pixel is simply the smallest controllable element of a given pixture, but then if you scale picture you start interpolating its data to ensure continuous representation. I'm sure that the scan data can do exactly the same: be represented as the only data of the image or with space between filled with interpolated information. :)

[@Kruglord]

This was a super cool video, it happens that I'm an engineer that works with archaeologists for just this type of stuff, specifically 3D capture and digital preservation of archaeological sites. It's very cool to see this type of technology being shown off, there's a whole world of different 3D measurement devices like this one.

[@hanniffydinn6019]

What other devices are there ? How do they compare to laser scanners ?

[@Kruglord]

+Hanniffy Dinn Well, there are many different types of laser scanners, but there are also a variety of devices that can be broadly classified as 3D cameras, as well as software packages which enable users to take regular cameras and, by taking many many photos, allow the user to recreate the scene digitally in 3D. This process is similar to the one used by the device in the video, but generally less precise, because things are less controlled.

[@simeon136]

+Kruglord We use a Leica scanner, Cyclone and Jetstream mainly for producing building elevations and sections in CAD, but the CYARK project fascinates me, and I'd like to get involved in more archaeological type work.

[@hanniffydinn6019]

simeon136 I've seen the exact same approach used to scan the surface of paintings. That also penetrated the surface using beyond normal eyesight wavelengths. I'm wondering if a laser scanner is detailed enough to produce point clouds for, say, something carved on a rock ?

[@simeon136]

+Hanniffy Dinn Yes, it is. My Boss scanned some carvings in a burial chamber in Wales and the results were very detailed.

[@teekanne15]

for all those wondering: Alex Pinz was born in 1958 in Vienna, Austria. He resceived the M.Sc. degree in electrical engineering in 1983 and the Ph.D degree in computer science in 1988 from Vienna University. In 1995 he received the habillitation in computer science from Graz University of Technology.

[@Obtaineudaimonia]

Simply beyond cool.

[@NikiHerl]

Wohoo, Austria! ^^

[@ihrbekommtmeinenrichtigennamen]

+Niki Herl Hört sich stark nach 'nem Wiener Akzent an :D

[@ozdergekko]

+Niki Herl -- ich hab auch gleich gedacht, dass dieses peinliche gestottere sehr nach Ostösterreich klingt. I tarat song: a Weana in Graz.

[@ZadakLeader]

I think the word he's looking for to describe the points is "Vertices"

[@hymen0callis]

+Vlad Ţepeş If there would be topology (triangles or polygons) they could be called _vertices_. In this modality it's normally just _points_.

[@luisdanielmesa]

+hymen0callis appropriate for tessellation

[@ZadakLeader]

I guess, but in order to make a 3d model you need polygons

[@Kruglord]

+Vlad Ţepeş In terms of most digital representations, points clouds are distinct from vertices. A point cloud (something I work with in my research extensively) really is just a collection of 3D points, XYZ coordinates, usually with some additional information included. That data is sometimes RGB colour, or sometimes just a single 'intensity' value indicating monochrome brightness. These points are stored in clouds, and are usually not dealt with on an individual level. A vertex on the other hand, though is also a 3D coordinate, usually exists on its own rather than in a set of vertices. They can have any number of additional pieces of information associated with them (name, height above ground, connectivity, etc).

[@panjmp]

+Vlad Ţepeş There are algorithms to construct 3d model from point cloud. But the data extracted from stereo image 3d reconstruction is point cloud (along with RGB).

[@panc8ke324]

I expect to see one of these on CSI....That'll be hard, as I don't watch it, but you get my point.

[@Shifticek]

+Panc8ke they'll probably be able to measure the precise position of atoms of any given surface

[@ajthemacboy8325]

+Brodne y And it'll function as a mass spec.

[@adiorthotos]

+Panc8ke ENHANCE!!

[@flaviomartinelli1803]

Valcamonica rules!! I live in there!

[@Computerphile]

+flavio martinelli A couple more videos on this project are in the pipeline, prepare to hear your home mentioned again! >Sean

[@andy_liga]

+flavio martinelli I'm italian as well and I've been there soooo many times, I will never forget being at the "Archeopark" as a child!

[@Triantalex]

ok?

[@GenaTrius]

Hopefully they can figure out a way to make this work for rock art on walls and ceilings, as well as floors.

[@simeon136]

+Gena Trius Then it ight be as useful as the existing technology already on the market. ;-)

[@jebus6kryst]

I have seen something similar to this in action last year on a site I was working on. They took 3D scan of the butt-hut we excavated. The 3D image was amazing.

[@Stefan_Payne]

Ah, he is an Austrian, that's why I thought I know that accent. That austrian accent always sounds pretty nice, even in english =)

[@hearueszueke6206]

Hey that's cool, I was by this professors lectures! Really interesting stuff.

Does it have an accelerometer or other type of measuring the orientation of the device itself? That should make the stitching of the surfaces more reliable when taking several pictures of the same surface. As an electronics and computer sciences student I find this thing to be quite interesting. :)

[@Seegalgalguntijak]

Is there a reason why not every LED is lit up at the same time, as can be seen in the slomo at 01:47?

[@Computerphile]

+Seegal Galguntijak I wondered, though it's probably a "rolling shutter" artifact of the way the camera works rather than the scanner - but I'll try to find out from Professor Pinz >Sean

[@Megabobster]

If it is just rolling shutter, look up SloMo Guys' video on how a camera works. Gavin explains it really well with some good slow motion footage.

[@ihrbekommtmeinenrichtigennamen]

+Seegal Galguntijak When you pause the video you can see that the boundary between lit and unlit LEDs crosses right through a single LED. Unless these are some very special kinds of LEDs it is not possible for them to be lit that way. So definitely rolling shutter.

[@Seegalgalguntijak]

Ihrbekommtmeinen Richtigennamennicht Right on, thanks mate!

[@Computerphile]

+Seegal Galguntijak Confirmation from Professor Pinz: "Yes, it must be related to the shutter in the camera. The flash is ultra-short. Otherwise, we can't produce the required intensity. The LEDs would overheat."

[@teekanne15]

might be useful for a fieldtrip in petrology, as often on some sights we are not allowed to take samples.

[@TheRealInscrutable]

that is really pretty cool. I wonder why they have to ask someone to make a sellable product though?

[@iamjimgroth]

I would love to have one for creating surfaces for games.

[@Kruglord]

+Jim Groth This devices seems to have a pretty limited field of view, so for the scale of a game you'd probably want to go with something like a terrestrial laser scanner. It creates the same soft of thing, but it covers much more area much faster. It wouldn't have as fine of detail or precise colour information, but it would be much faster and easier for large areas.

[@oOfretlessOo]

+Kruglord I was thinking upscaling of small textures like sand formations and such to make unique environments.

[@Kruglord]

+oOfretlessOo Oh, that would be interesting. I thought you were thinking of making a game environment out of existing architecture or physical locations.

[@abeedhal6519]

+Tasty Mixes interesting, can you tell me where you got that information from?

[@BubuSnow93]

+Jim Groth They used a similar technology for the game "The Vanishing of Ethan Carter"

[@hathejoker]

Cool setup

[@DarkAngelEU]

I'd love to make a portrait with this camera, just to see how it looks on living objects :)

[@KennethMeyerson]

Very, very cool!

[@Rififi50]

I have to say, I was wondering whether that was a German accent but I couldn't pinpoint it. As soon as he mentioned Graz, I just couldn't get the Austrian accent out of my head. (^_^')

[@istoOi]

+Rififi50 if someone sounds like Arnold Schwarzenegger, it is probably an austrian accent :)

[@0101dyon]

+Rififi50 It's dutch I believe. I can hear the dutch. Could also be belgian. Im always annoyed by the accent us dutch people have. It's soooo bad... his isn't so bad actually, but can still hear it.

[@LukeSumIpsePatremTe]

+Dyon ”Diwany” de Kok It can be a little hard to understand sometimes. But it was okay. I've heard that native english speakers have more difficulty with accents than some of us, for whom english is second language.

[@Rififi50]

istoOi I basically never hear Arnold Schwarzenegger speak English (^_^) FYI, Arnold actually comes from a suburb of Graz.

[@0101dyon]

***** Ahhh. That does explain the french-ish twist I heard. Thought it as belgian, but Austrian does make more sense

[@MBaadsgaard]

How do you correctly recreate the normal map here? Is it from the stereo imaging or is it from the difference between the two image sets, or how?

[@recklessroges]

Today on DragonsDen^WComputerphile.. I wonder how long before this is used for crime scene photography. I'm sure there are other applications as well.

[@Kruglord]

+Reckless Roges Similar technologies are already in use, actually.

[@silmarian]

Darn, it doesn't look like they have the models available for download on their site.

[@insu_na]

That is a *real* German accent (the guy is from Austria). I can't think of any other use-case for a device like this. I thought agricultural sciences, but stereo isn't quite enough for that.

[@simeon136]

If you want to see truly impressive 3d rock art scanning, look for the work done by Dr George Nash of Bristol Uni and Andrew Beardsley of Terra Measurement at Barclodiad y Gawres on Anglesey.

[@avro549B]

V 2.0 will be an add-one for a smart-phone.

[@toxicore1190]

did i hear point cloud? whats about eucledeon

[@borzan007]

Is there any way to get email contact from this man ?

[@vizionthing]

This was already awesome .... until he tells us its controlled by a Rasberry Pi .. that ups the awesomeness by a factor or two!

[@b4ux1t3-tech]

Not a factor of pi?

[@vizionthing]

hmm I should have phrased that better :)

[@mikosoft]

+vizionthing I understood that Pi only gets the JPEGs and sends them to the tablet

[@Folopolis]

+vizionthing I thought it made it seem somewhat amateurish.

[@vizionthing]

Folopolis Armatures built the world, don't under rate them! I happen to be a big fan of the Raspberry Pi having followed its development all the way to market, it nice to see them being used for more than just teaching.

[@cbrpnk]

Why no example?

[@Kruglord]

+cbrpnk It probably needs to be published in a scientific journal before it's shared with the public. There are a lot of reasons why they wouldn't want to release that sort of stuff before it's been peer reviewed.

[@cbrpnk]

Kruglord It makes sense.

[@AlastairToft]

Downing College?

[@TechyBen]

Lasers. Lasers fix everything*. *That a personal jetpack or flying car cannot.

[@RamboToReality]

nice

[@Durakken]

Why couldn't you use it along with the sphere camera with crime scene investigations or anything that you need to get as many details of as possible out of? or just regular art? Does it really have to be that finely calibrated... From what I understand the way it works is by knowing where the LEDs are in Reference to the Camera lenses and in the case of multiple images, a static reference of some sort to them, but seems to me that isn't actually an issue since you can fix that with image matching software or something as simply as if you have a cable running over and area and are moving along it, the speed of tha camera's movement should be enough to match edges... likewise if you know the direction and speed, i see no reason why you couldn't use this with something like a drone which has to record and pass on that information anyways...

[@EnricoPiazza]

+Durakken - knowing the speed on a drone is actually very difficult. Even if you somehow work out the drone's speed, it would still not be precise enough for that kind of fine measurement.

[@superdau]

+Durakken 3D models of the ground/environment taken from a drone is being done quite alot already. Speed and direction is pretty irrelevant, the images are just aligned by features. Resolution and/or accuracy in the range of a few cm is quites easy to achieve nowadays. But this scanner has a different purpose.

[@TheGreatSteve]

This is the opposite of ISIS.

[@therealpanse]

just asking... I'm a native Gaerman speaking. Is it just me, or is he Austrian? Just wanna know if I can hear his accent through another language. Maybe Bavarian, but pretty sure he's Austrian. The way he pronounces (or better not pronounces) consonants sounds very German and at the same time very smooth like someone from the alp region.

[@therealpanse]

alright, just watched the rest of the video and he said he' working at Graz, so I was right. You can't unhear German languages from English...

[@ihrbekommtmeinenrichtigennamen]

+therealpanse Ich hab am Anfang an Wien gedacht, aber Graz hört sich auch plausibel an. Und ja: Man kann Akzente in der Art definitiv raushören. Ich selber spreche sehr wenig Englisch (Internet ist halt mehr lesen, schreiben und hören) und hab deshalb einen extremem Akzent beim Sprechen.

[@moatl6945]

+therealpanse I'm a native German/Bavarian speaker. I think the main characteristic is more in the pronunciation of the vowels. Especially the pronunciation of the vowel »a« is very specific for a person coming from the East-Middle-Bavarian area (Vienna, Graz, Linz,…). At least it was the thing it made my east »ringing«. :) If you compare the pronunciation of the same words as spoken in the West-Middle-Bavarian area (Munich, Ingolstadt, Passau,…) in the western areas you can hear two different sounds for »a« (a »dark a« and a »bright a«), where in the eastern areas you can only one sound (a »brighter a«).

[@superdau]

+Martin Steindl For me (Austrian) it's more the intonation that gives it away. In this case the melody of the sentences just made me think "ah, a Weana" ;) . And there's one thing, that makes "german english" much easier to recognize: the "ze". For some reason Germans very often pronounce "the" as "ze" or "se" (which the native english speakers joke about), while in Austria I've only heard "te".

[@Tenus123]

Yey!

[@markwilliams5654]

the should take it to Göbekli Tepe and find our real human history

[@TechXSoftware]

haha, very expensive equipment to take pictures of rocks lol

[@psp09ful]

What...

[@claushellsing]

Linux powered!!! :) yeah!!

[@Roshy1]

stuck at 304 views? I thought this was fixed.

[@KyleHarrington1986]

What this needs is a nonsensical abstract art video that makes no mention of 3d Pitoti project or the technical requirements involved...

[@seanski44]

😉

[@hannovb5379]

first1!!!!!

[@Erin-bc8ic]

first!

[@artifactingreality]

I would never pay 5000euros for that. give me my tax money back

[@EnricoPiazza]

+artifactingreality - You are clearly not an archeologist.

[@viennaorange4122]

comment