Slide 1
Slide 1 text
3D4EM
Storage, maintenance and dissemination
Hugo Ledoux
TU Delft
3D4EM kickoff meeting
Rotterdam, 2015-01-14
Slide 2
Slide 2 text
No content
Slide 3
Slide 3 text
No content
Slide 4
Slide 4 text
No content
Slide 5
Slide 5 text
No content
Slide 6
Slide 6 text
No content
Slide 7
Slide 7 text
Storage
Slide 8
Slide 8 text
What 3D objects do we have?
8
• 3D buildings in LOD0/1/2 → PostGIS
• vegetations (in diff LODs?) → PostGIS
• terrain
• 3D roads
• canals/water
→ 15M Top10NL objects
→ 1B+ BGT objects
→ 640B AHN2 elevation points
TIN in a DBMS = complex
}
}“easy”
Slide 9
Slide 9 text
Massive TINs in a DBMS
9
Slide 10
Slide 10 text
10
Slide 11
Slide 11 text
11
Slide 12
Slide 12 text
12
Slide 13
Slide 13 text
13
Slide 14
Slide 14 text
14
Slide 15
Slide 15 text
15
p
a
b
c
d
e
f
Slide 16
Slide 16 text
16
p
a
b
c
d
e
f
Slide 17
Slide 17 text
17
p
a
b
c
d
e
f
star(p) = abcdef
Slide 18
Slide 18 text
and we store the star for every vertex
18
Slide 19
Slide 19 text
and we store the star for every vertex
19
Slide 20
Slide 20 text
and we store the star for every vertex
20
Slide 21
Slide 21 text
and we store the star for every vertex
21
Slide 22
Slide 22 text
Stars in a DBMS
ID
x y z
star
1 3.21 5.23 2.11 2–44–55–61–23
2 5.19 29.01 4.55 7–98–111–233–222
3 22.43 15.99 8.19 99–101–73–23
... ... ... ... ...
5674 221.19 15.23 37.81 309–802–793–1111
Advantages:
1 Only one table with
id x y z star
2 No spatial index needed: fetching of triangles based on
“walking”
3 Star column need not be filled (⇠ Simple Features)
4 Local updates are possible (insertion and removals)
5 Ideas are readily extensible to 3D for storing manipulating
tetrahedra
Slide 23
Slide 23 text
Research objectives
23
1. massive TINs (1% of 640B = 6.4B points)
2. constrained TINs
3. TINs with vertical faces and ‘overfolds’, ie
having a full “3D TIN”, not only a 2.5D TIN
Slide 24
Slide 24 text
some preliminary results
24
Storing and analysing massive TINs in a DBMS with a star-based data structure 17
degree
points duplicates triangles convexhull avg max
AHN2 281 884 687 214 050 563 768 199 1 173 6.00 63
serpent 3 265 110 17 584 6 494 998 52 6.00 39
msh 283 213 392 0 566 426 669 113 6.00 141
Table 2 Details concerning the datasets used for the experiments; convexhull is the number
of points that are on the boundary of the convex hull of the dataset.
Storing and analysing massive TINs in a DBMS with a star-based data structure 19
star structure triangles SF
table index total table index total
AHN2 28 GB 4.8 GB 32.8 GB 64 GB 29 GB 93 GB
serpent 325 MB 56 MB 381 MB 746 MB 329 MB 1075 MB
msh 28 GB 4.8 GB 32.8 GB 64 GB 29 GB 93 GB
Table 4 Size of the tables and the indexes in PostgreSQL for the datasets.
of triangles is around twice as large as that of points), the main cause is the
building of the GiST spatial index of PostGIS, which took 20 times more time.
Slide 25
Slide 25 text
some preliminary results
25
Storing and analysing massive TINs in a DBMS with a star-based data structure 17
degree
points duplicates triangles convexhull avg max
AHN2 281 884 687 214 050 563 768 199 1 173 6.00 63
serpent 3 265 110 17 584 6 494 998 52 6.00 39
msh 283 213 392 0 566 426 669 113 6.00 141
Table 2 Details concerning the datasets used for the experiments; convexhull is the number
of points that are on the boundary of the convex hull of the dataset.
Storing and analysing massive TINs in a DBMS with a star-based data structure 19
star structure triangles SF
table index total table index total
AHN2 28 GB 4.8 GB 32.8 GB 64 GB 29 GB 93 GB
serpent 325 MB 56 MB 381 MB 746 MB 329 MB 1075 MB
msh 28 GB 4.8 GB 32.8 GB 64 GB 29 GB 93 GB
Table 4 Size of the tables and the indexes in PostgreSQL for the datasets.
of triangles is around twice as large as that of points), the main cause is the
building of the GiST spatial index of PostGIS, which took 20 times more time.
Slide 26
Slide 26 text
Maintenance
Slide 27
Slide 27 text
No content
Slide 28
Slide 28 text
visible errors
Slide 29
Slide 29 text
Dissemination platform
Slide 30
Slide 30 text
www.3d4em.nl
Select area:
Choose format:
Choose format:
Download 3D TOPNL
CityGML 2.0
VRML
ArcGIS
Applications:
Noise modelling
Wind modelling
Flooding
CityGML 3.0
size:
voxel (netCDF)
5m x 5m x 5m
Specifications:
2.5 TIN
trees
buildings LOD1
buildings LOD2
Save
Slide 31
Slide 31 text
A typical Dutch landscape
31
hill tree house
sidewalk
Slide 32
Slide 32 text
2.5D terrain + separate 3D objects
32
Slide 33
Slide 33 text
2.5D terrain + separate 3D objects (LOD1)
33
Slide 34
Slide 34 text
3D terrain + separate 3D objects
34
Slide 35
Slide 35 text
3D DSM (digital surface model)
35
Slide 36
Slide 36 text
2.5D DSM
36
Slide 37
Slide 37 text
3D grid (voxel)
37