into this venture what this point while
modeling is about we founded seven years
ago the ETH together with professor
Greer Egham a laboratory bridging
engineering and architecture it was
supported by the Swiss National Science
Foundation and the school board the ETH
School Board and what happened is that
we transferred a technology that was
privy to engineers and surveyors and
brought it into the realm of modeling
and the way we argued with the Swiss
National Science Foundation is we
basically said well there’s no such
thing as a landscape modeling laboratory
a DTH but not even in Europe or probably
worldwide we’d like to bring this sort
of large-scale modeling from the virtual
world into a reality and make it
accessible to younger generations of
researchers and students so that’s
actually what we did what you’re going
to see is going to require a little bit
of a gymnastic to understand how this
sort of visualizing occurs but actually
what you’re looking at in this picture
is a still picture of a model it’s a
gigantic digital model showing the
entire Alpine chain ranging north south
over the Guittard
before we start I think that we should
define first what a point cloud is made
of it’s actually made of millions if not
billions of pixel points the pixel
itself is a recent word in the history
of mankind it appeared in 1967 at NASA
actually in the Jet Propulsion lab
Billingsly when they satellites and and
spatial missions were sending back
electronic image to earth and it simply
means picture element it’s the smallest
unit visible unit on a digital image or
screen it goes back actually to 19
century theories German person called
Paul nip cough first talked about a
built Punk’d the basically that’s what
the entire cloud is made of so what is a
cloud it’s basically a set of points
that carry information each point
carries in a way a position relative to
its neighbors and the black space in
between is simply no information and as
I said until recently these data sets
these point clouds were used by very
serious engineers civil engineers mostly
to draw precise technical plans and
elevations on construction sites like
dams and other things and the image
itself while the model itself was never
made fully the information was extracted
and 99% of the point cloud information
was thrown away what happened with this
laboratory is we bridged the two fields
engineering and creative architecture
and kept the entire datasets as they
were and started working with these
unwieldy models that count billions of
points basically each point carries the
famous GPS coordinates the X Y Z that
assesses slant latitude longitude and
elevation and you have within this
family of point clouds these so-called
control points that anchor the entire
cloud in precision and we’re talking
about millimetre in terms of height or
width the position is within millimeter
precision these control points are
actually riveted to a series of
satellites that go around the earth and
really give that fixed position the
control point also helps to bring
different sets we use drones we use
terrestrial lasers we use lidar mobile
cars that register elements and these
control points allow us to bring things
together so much for the technical
explanation I think the best way to
explain that is to start again with the
point within a point cloud so that you
can’t stand the sort of phenomenon that
I call trance duck
scale what happens is you can go from a
minut scale of an object to the very
broad scale of a territory seeming
seamlessly and that’s quite
extraordinary so let’s start our journey
here we’re going to see a pixel point
retracting into the virtual digital
space of the new net tunnel this is
actually a data set that was given to us
by the ARP transit of the entire new
high speed rail tunnel in Switzerland
the longest in the world and here we’re
retracting out of the south portal which
is in BO do very slowly and we’re going
to pull out at the scale of an object
and as we go up in the air we’re going
to literally start seeing and this is
all a point cloud data set we’re going
to start seeing the model in a much
bigger picture of the environment here
you even see raindrops falling on the
tunnel entrance
they were scanned the forests the
mountains adjacent to that etc so what’s
extraordinary with the point cloud when
you use it as a general data set is it’s
a virtual model it’s a still image but
it has extraordinary precision this is
just the section of the model that we’ve
been scanning at the eth over the last
seven years of the landscape modeling
and visualizing laboratory what you see
this sort of strange little streak on
the upper part of the the plan picture
is actually the datasets that we took
they’re embedded in a Swiss topo dataset
and what you see underneath is the
section through the mountains with that
high-speed tunnel going through if you
look at the screen you also see that her
thin line in the middle of the picture
which is literally the tunnel that you
see in semi-transparency through the
pixel points of the model this is the
tunnel you all know it it’s been very
highly publicized since its inauguration
last June it is quite unreal one really
doesn’t know where the one is under the
ocean or under the mountain the
relationship to the landscape has been
in a way lost here and of course when
you look at what has happened with this
highly technical sort of device
is that you can cross the Alps in
roughly 20 minutes it used to take two
weeks when Goethe went over the Alps and
basically what you see in gray is the
entire mountain landscape having
disappeared what’s good though is that
the Swiss train company the SBB has
provided for us a virtual compensation
they have within the trains as you cross
the tunnel monitors that talk about your
new horizon it’s labeled underneath each
screen and probably it’s the absence
thereof that means that the new horizon
appears on a virtual screen so what
should we say about that I think we’re
going to go back to the point cloud to
try and understand how it actually works
as a model and as a piece and I think
the best way to designate that is to
explain that the model is like a crust
here we’re just south of the Gothic
going down we’re monitoring the earth
crust it’s really showing in the most
intricate details the old tremolo wrote
in the new tremolo Road it’s almost like
a big cake a digital cake which is empty
underneath it’s just really showing the
surface in the most intricate detail
what you’re going to start seeing now
though is the tunnel below that line
that appears in transparency which is
actually the autobahn tunnel the first
tunnel that we’re going to see and it’s
appearing in transparency through the
cross the second line above that looks
like a cloth hanging line is actually
the nail tunnel which is further east
and as we cut through that model as if
you’re cutting through a piece of cake
we see with exact geographic position
the depth of field with the two tunnels
below the earth crust that’s quite a big
will during image and that’s where these
point cloud data sets are actually
producing a sort of revolution of sorts
again the point cloud is a surface that
is geo-referenced and that really shows
an extraordinary amount of texture here
you see the texture in front of the
terrain but you also see the
transparency of the entire mountain
chain behind it goes all the way to the
Mater hole on the far left this is part
of the model again and here we see the
journey
north-to-south so I’ll just let you
travel we’re going to go through the
portal that we just saw before what’s
extraordinary with this vision is that
we actually come down and within the
recognizable human scale we see things
in very very precise positioning I think
here we’re going to get to the portal
that we just left a few minutes before
and what you’re going to see very
quickly is the portal and then the
double pipe of the train line veering to
the right under the mounting crust and
then shooting into the mountain itself
transparently here we’re going up the
hill towards the Guittard
and then we’ll come down over the
devil’s bridge below you see the
motorway tunnel in transparency that’s
really what’s absolutely fascinating
so it’s actually a massive scale digital
model that can be used here we’re
arriving on the Gothard again and then
we go down again with the mote ultra ban
bridge underneath and magician and fluke
The Devil’s Bridge the well-known
Devil’s Bridge on the bottom of the
picture etc going through we’ve been
deploying techniques like drones before
they were celebrated that was 10 years
ago now they’re sort of ubiquitous we’ve
been using the lidar system we even
reserved the Ultra Bank tunnel for
ourselves close it off and measured it
we even have been going to very cold
places this is Matthias scanning the
Guittard what I think is really
important to understand is how this
really works and what its applications
are for the future it’s not just a
nice-to-have model of the Swiss Alps
here we can literally work as engineers
sexually we’re going to cut through the
infrastructure that pours out of the
Alps in Goshen and here you see a very
very precise section showing how the
motorway the railway bridge etc works
and the applications there have been
multiple we’ve worked in Southeast Asia
and Jakarta monitoring flood problems in
slums we’ve been rebuilding artificial
islands in Venice with the same
technology we’ve been working in Israel
and Jordan on the naira your insight to
document archaeological elements and
basically what we’re able to do with
this type of modeling is not just map
the Alps but actually simulate flow
simulate new projects and use it almost
as a tool of participation we’ve trained
hundreds of young researchers and
students for it and I think that’s
really the goal of this laboratory so
basically the entire team behind the
guitar project which was shown
by the way in the Venice Biennale two
years ago and will be shown again at the
ETH Center at the beginning of next year
in June is actually the result of a huge
amount of collaborations I thank you for
your attention
[Applause] [Music]