Upload
others
View
12
Download
0
Embed Size (px)
Citation preview
19/11/2018
1
P1
P2
Engineering Perspective of Geo-Spatial
Intelligence in Smart City
Huang Qian
19 November 2018
19/11/2018
2
P3
What is a Smarter City?
Physical
Energy
Information
Emission reduction
GreenGreen
Biodiversity
Livable
Networking
Digitization
Intelligent
SensorSensorControllerController
ControllerController
SensorSensor
Ecology
Wisdom
P4
What Computers Do: Model, Connect, and Engage
参与(Engage)
联结(Connect)
建模(Model)
Butler LampsonA.M. Turing Award Laureate
Technical Fellow at Microsoft
Corporation
Adjunct Professor of Computer
Science and Electrical
Engineering at MIT
����������� �������� ����������
���������������!��� ��#��"��$%
�&'(�*)'(�+,-�� .���/ 2 ��,0 ��� .���/%
AI
Big
Data
5G
IoT
Digital
Twins
19/11/2018
3
P5
Real World
Information World
People
Internet of
People things
Internet of
things
Object
ID
Virtual
Represent
Control
Mathematical
Model
Mathematical
Model
Connection, more and faster
202513456789:;<=53.8>?@A5G56789:<=39.3>B ——C20181A45GDEFGHIJKLMNOP
P6
Engage, interactive and automatic
Depth perception
Edge calculation
Time-space migration
Collaborative control network
19/11/2018
4
P7
Model, Geo-Spatial Big Data
Point
Q0-DR
Line
Q1-DR
Polygon
Q2-DR
Body
Q3-DR
Mutil-D
Temporal
Spatio
static
dynamic
MacroMicro
Simple
Complex
transportatio
n network
underground
pipe network
Function
Zoning
Land use
BIM
Mutil-
Domain
Intgeration
City
General
Indicator
Small
Big
Agent Based
Discrete Event Simulation
System dynamics
P8
Geo System
Geo-Spatial
TechnologyGeoControlGeoMonitor
GeoModelingGeoModeling
Data
Model
Math
Model
GeoDesign
sensor controller
IoT
BIM
Cooperative object
network
Earth Observation
Network
Evolution of new Geo-Spatial Technology
19/11/2018
5
P9
Geo-Spatial
Intelligence
ST(discover)
UV(foresight)
WX(insight)
YZ(organization)
[\(tools)
]^(process)
_`abcdefghiffjdkilmfn
T
op(system)
qr(E
nv)
Geo-Spatial Intelligence Structure and Mechanism
st(fusion)
u�v
w ����
xyzy{|}~�|������|{y��
�}��������{y�z
��
��{����{�
P10
�|z �� ��|� �|y�{��|����|�|z����{
xy��y��}{y�� y� �yz��|�����{���{y�� �|�|z����{
�|�� ����y{y�� y� �y�}�y����{yz|{y��
������������� �� ���� ��
¡�|¢ £|�y� �� ��|�¤y�����|{y¥|{y��
¦§�§������ �� ¨�©�§ ª§«� �
Can we reform traditional transportation?
19/11/2018
6
P11
¬®5������� ¯��/0%
°±²³´µ¶·¸)
¹º»¼½¾¿ÀÁÂÿÄ
QÅÆÇÈÉÇODÇÊËÌR
ÍÎÏнÑÒÓÔÕÖ×ÔÕÂ×ÕÔ¿Ä
(ØÙÇÚÛÜÝÞÇßàáR
âã�äå� �0%(æç´èé)
Ëê
ë¿ìíîÔïÃðñòó
ôê
õíï×ÁïÃ
ö÷øù
úû
üà
ï××ÔÁýÕ׿þÿ��
System Analysis
P12
Big Data Processing and Sharing Platform
Scientific
Support
Industry
Revolution
Smart Operation
& Maintenance
Community
Service
�������������
� ����������������
��� ����������� ����������
������� �������������� �����
Transformation
& Upgrading
��� !"#$%&'$"&($)*%*�"
�$)�""�*%%�")$+ ($%*,"
��� -�"%&./)&*�"
0�*"&$"�")$ 0�"�,$'$"&
12$.�&*�"�30�"�,$'$"&
0�.4$& -.$ *&
Data
Layer
Tech
Layer
App
Layer
Tech Stack
19/11/2018
7
P13
Case 1,Remote Sensing Recognition
P14
Algorithm, Generative Adversarial Networks
567689:;8
<=>?8=@=79:;8 A;8BC
DEFG
HIFG
� 生成器:生成以假乱真的图像� 鉴别器:尽可能鉴别真假� 当鉴别器无法区分真假时,模型训练完毕。
JKDE
19/11/2018
8
P15
Model Fusing
Multi-scale fusing
多框架︑多参数︑多模型
噪声少、连续性好
LM
LM
LM
NL
P16
Post-treatment, Rattlesnake
利用自动生长、“响尾蛇”等算法,实现路网的自动连接
OPHQ
RSTUV
OPHQ
19/11/2018
9
P17
Building Recognition
P18
Xiamen
19/11/2018
10
P19
Faster
以25平方公里的区域为例,人工勾画路网需8小时/人,利用机器学习自动提取,
结合半自动化的后处理,仅需半小时/人
W
X
Y
Z
[
\
]
^
_
`abc defg
hijkhl
P20
mnopqr
stqr
Better
19/11/2018
11
P21
养护 uvwx uvyz
Algorithm in Smart Maintenance
P22
Traffic Volume Forecast
{|à}~����
���� ����� ������ ��������� ���� ������ �� ������
�����
���� ����
���� ����� ������ ��������� ���� ������ �� ������
�����
���� ����
���� ����� ������ ��������� ���� ������ �� ������
�����
���� ����
��������
� ¡¢£¤{¥
¦§¨©ª«
¬¦®ª«
¢£¯ÚÉ°
±²ÝÞ¬¦¹³ÝÞ
´µ¢£¶·¸¹
19/11/2018
12
P23
º»�¼ º»�¼ º»�¼ º»�¼ º»�¼ º»�¼
½¾¿¾ ½¾¿¿ ½¾¿½
½¾¿À ½¾¿Á ½¾¿Â
交通量影响因子
交通量
recurrent neural network(RNN)
Synthetic control method
P24
山西省全路网交通量预测 福建省拟建高速公路交通量预测
Traffic Volume Forecast Platform
19/11/2018
13
P25
主体(subjects)
工具(tools)
ÃÄ(algorithm)
Åã(software)
ÆÇ(system)
ÈÉ(customer)
ÆÇÊIË(developer)
ÌÍÎÏ(domain expert)
:ÐÑÒÏ(data scientist)
ÓÔÕÖË(tools provider)
:ÐÕÖË(data provider)×Ø
(individual)
ÙØ(group)
ÚØ(ensemble)
�������%
ÛÜ(Information)
(entity-semantic-structure)
ÝÞ(Knowledge)
(relation-rule-pattern)
ßàáâ(individual
portrait)
ãä(heterogeneo
us)
åàáâ(group
portrait)
æäçä(classificatio
n;
clustering)
èàáâ(ensemble
portrait)
éÚêë(aggregation
structure)
:ÐìÝ(data
sense)
ÛÜÕí(information
extraction)
ÝÞîïKnowledge
condense
客体(objects)
Computer
Engineer
Urban
Planning/
Smart CityArtificial
Intelligence/
IoT….
Engineer Framework for Smart City
P26
ðñòóGeo-System
ôõö
Controler
÷øö
Sensor
ùúûüSystem Output
ýþûüMeasured Output
ùúûÿSystem Input
ýü��Measured Error
�����
Reference
GIS
GeoModeling
GeoDesign
GeoControlGeoMonito
r
Feedback
Geospatial Intelligence Helps City Smarter