Geoscience Reference
In-Depth Information
Table 1
The compression ratio (near-lossless) with quantization option set to true
Original
(MB)
Deflate
zip (MB)
Fast-
Infoset
(MB)
LZMA
(MB)
Bzip2
(MB)
CitySAC
? Deflate
(MB)
CitySAC
? LZMA
(MB)
Commercial_
Building3C6.xml
0.836
0.086
0.469
0.073
0.079
0.072
(91.39 %)
0.056
(93.3 %)
National Audit.xml
8.594
0.948
4.132
0.696
0.869
0.723
(91.59 %)
0.561
(93.47 %)
Putrajaya_
Mosque.xml
10.928
1.56
4.81
0.992
1.226
0.98
(91.03 %)
0.782
(92.84 %)
Seri Gemilang
Bride.xml
27.454
3.612
12.352
2.714
3.519
2.59
(90.57 %)
1.94
(92.93 %)
Table 2
The time consumption for different techniques with default setting (Using Stopwatch
library)
Input \ Techniques
Deflate
zip (sec)
Fast-
Infoset
(sec)
LZMA
(sec)
Bzip2
(sec)
CitySAC
? Deflate
(sec)
CitySAC
? LZMA
(sec)
Commercial_Bulding_3C6.xml
0.2
0.5
0.3
0.1
0.3
0.4
National Audit.xml
3.8
3.6
4.5
1.2
1.6
1.9
Putrajaya_Mosque.xml
3.5
4.2
5.8
2.2
3.1
4.1
SeriGemilangBride.xml
9.8
11.5
15.2
4.5
10.5
11.2
Since CitySAC use dictionary encoding, all occurrences are indexed and stored.
Furthermore, geometries are built in chunk of 65,000 face sets where each face
represents a polygon in CityGML. The main idea of the encoder is to follow the
XML original structure so query-able capability is retained and compressed con-
tent is retrievable. The encoder defines each representation as a symbol denoted in
16-bits.
The compression ratio is calculated with the following equation:
100 %
CR
2
¼ 1
size of
ð
compressed file)
size of
ð
original file)
WinZip is used as the benchmark, and this encoder aims to perform better than
direct 7-zip results in terms of timespan and compressed size and used 7-zip the
final compression stage. It is worth mentioning that the encoder achieved a
35-50 % better compression rate compare to WinZip, and 20-30 % better com-
pression rate compare to direct 7-zip (See Table
4
). In near-lossless scheme, better
compression ratio is achieved (see Tables
1
,
2
,
3
). The environment for encoding
is shown in Table
4
.
