smart spatial data management @ ordnance survey
TRANSCRIPT
THE FUTURE IS DIGITAL
Smart Spatial Data Management
Debbie WilsonEnterprise Data Architect Oracle BIWA Summit 26th-28th Jan 2016
[insert protective marking - see QSP 032]
Apple Google, Uber and Car Manufacturers are making global maps but are not National Mapping Agencies
WELCOME TO 4th INDUSTRIAL REVOLUTION
Big Data tech & BI Data visualisation are encroaching into the GIS space
UNDERPIN SYSTEM OF SYSTEMS FOR SMART DECISION MAKING
Finance & Insurance
Energy & Infrastructure
Health &Emergency
Services
National & Local
Government
EnvironmentLand & Property
Leisure
Intelligent Transport
Oracle BIWA Summit 26th-28th Jan 2016
Data that underpins £100 billion of Great Britain’s economyHelping Great Britain generate £400 million in new revenues every year
Core Reference Geography
HISTORY – PRE-DIGITAL (1791– 1971)
1791
1950
1971
Major-General William Roy carries out the first
triangulation survey
Brigadier Martin Hotine leads a highly accurate re-triangulation of Britain.
Introduction of Digital Mapping to large scale map production
Oracle BIWA Summit 26th-28th Jan 2016
OS has been digital for over 40 years!
A History of OS
HISTORY - GOING DIGITAL (1971 – 2011)
1971 1995 2000 2003 2008 2011
Improvement in surveying tech
Civilian Organisation1980s
1st in world to complete national large scale digital
map OS LandLine(CARTOGRAPHIC)
Launched OS MasterMap – seamless intelligent object dataset
“more than a map”(CARTOLYTIC)
Operational use of GDMS: disconnected large scale editing system allowing simultaneous
editing by internal & external staff (field/off-shore)
Oracle BIWA Summit 26th-28th Jan 2016
[insert protective marking - see QSP 032]
In 2008 we had an integrated and seamless large-scale geospatial product OS MasterMap But it was underpinned by:
• separate production systems• tiled maintenance regimes• legacy of digital cartography – transition to geographic information incomplete• Inflexible and disparate systems limiting our ability to extend our data specifications
GEOSPATIAL DATA MANAGEMENT SYSTEM (GDMS)
GEOSPATIAL DATA MANAGEMENT SYSTEM (GDMS)
NOTE: Oracle components have
been upgraded!
Oracle BIWA Summit 26th-28th Jan 2016
460 MILLION OBJECTS SURVEYED
~ 0.2 - 0.5 million individual objects surveyed/verified per month
Topographic layer: 245,000 km² of Great Britain
320 internal & 880-1,200 external contractors continuously updating topographic layer (24/7)
Oracle BIWA Summit 26th-28th Jan 2016
[insert protective marking - see QSP 032]
GDMS was meant to scale to support new products – but this reality has quite happenedImplementation conflated many processes into single long transaction increasing costs to changeResulting in need to establish additional maintenance data stores to support new productsGDMS is going to be re-developed with an emphasis on increased automation, use of new sensors technologies and co-creation/integration with third parties in order to gain efficiencies and remain competitive
GEOSPATIAL DATA MANAGEMENT SYSTEM (GDMS)
Sites
Urban Path Network
HISTORY – AUTOMATION ERA (2011 – 2015)
2011
2013
2014
2015
VectorMap District (VMD) Beta created from prototype automatic generalisation
system launched kicking off Multi-Resolution Database Programme (MRDP)
Automatic derivation of OS Terrain 50 and Terrain 5 from
detailed height store
Automatic calculation of Building Heights from surface and terrain
model for large scale topo objects
1. MRDP Generalisation Information Engine (GenIE) operational production of VMD
2. Automated Object Recognition to classify field boundary objects (hedges, walls)
Oracle BIWA Summit 26th-28th Jan 2016
AUTOMATIC GENERALISATION (MRDP)
Large scale buildings (1:1,250) District scale buildings (1:15,000-1:30,000)
‘To improve the currency and consistency of current products and allow simpler creation of new derived products that meet changing business and customer needs.’
Oracle BIWA Summit 26th-28th Jan 2016
[insert protective marking - see QSP 032]
LOGICAL ARCHITECTURE
Stor
age
Tier
Data Validation, Quality Assurance& Cartographic Fixing
Product Publication
Data Enhancement & Generalisation
PHYSICAL ARCHITECTURE: GENIE
Change Manager &
Job Aggregation
1Generalise+ Rules
+ Roads Code+ Buildings Code
1Validate+ Rules
Rounding
Join Features
Content Product 1Workflow
1Plan
1Exchange
1Transact GoLoader
GoPublisher
Editor
Job starter
1Worklist
Publication Orchestration
Change Manager UI
Publication UI
Raster Service
Vector Service
Provisioning Service
Large Scale
Sources
[insert protective marking - see QSP 032]
AUTOMATIC GENERALISATION – VectorMap District
To generalise/validate content & product: • 12 Generalisation services - 4 nodes • 6 Validation services – 2 nodes • 6 days: 3 content & 3 product• 24.5 million features
Raster (colour & b/w) and vector map production (GML/SHP): 6590 tiles in 2 days
End-to-end automated process: < 10 days Can increase product publication schedule from bi-annually to 2
weekly with monthly currency
Validation failures: 0.002% failure rate• 400 features during content generalisation• 250 features during product generalisation• 8 man days to fix (2 days elapsed)
AUTOMATIC OBJECT IDENTIFICATION - HEDGESAutomatically classify Topo Lines as Hedges, walls, trees around Rural Land Register parcels
AUTOMATIC OBJECT IDENTIFICATION - HEDGES
Classified Land Cover Content Store714 M polygons – national predicted at 2 Billion objects 1.8 TB)Land Cover types: tree, scrub, grass, man-madeAutomatically classified using Imagery, Height and Topo
Hedge Product Store11 M Hedges – national predicted at 16 MOutput full product set as gml tiles in 21 minutes
Image Store1 x 1 km at 15cm resolution268,000 tilesDatabase size 79.8 TB
FUTURE– “SMART” ERA (2016 - Future)
Oracle BIWA Summit 26th-28th Jan 2016
2016
2018
2020
2025
Automatic integration of local authority Address and Highways data
Cloud Infrastructure end-to-endRapidly integrate data from multiple
sources to delivering enhanced content
Real-time consumption of content (Data-as-a-service) & analytics/BI
Digital Built Britain (BIM/Smart Cities)Intelligent Transport Systems
Contact: [email protected] Oracle BIWA Summit 26th-28th Jan 2016