CityGML turns structural components into an index

Report
CityGML+ARML Workshop
Discussion 1
Introduction
Alex Karman, CTO and Co-Founder
Revolutionary Machines, Inc.
Using CityGML As a Universal Index of
the Physical World
Our goal: a visual, browsable, 3D index into
everything that intersects with the urban
landscape
– Groups, roles and responsibilities
– Assets and resources
– Workflows and processes
– Analytics and intelligence
Our Approach
• Use CityGML not as data, but as an index
– Don’t modify CityGML to add more properties
– Use XLINK hooks to external RDF data sets
– Use SPARQL and Xquery to browse
• Use the Semantic Web to add context to the
urban environment
• Use Augmented Reality to filter locationspecific and role-specific information to actors
Benefits of this Approach
•
•
•
•
•
Fast federated search and retrieval
Interoperability over multidisciplinary data sets
Cooperation without coordination (“serendipity”)
Flexibility and Agility
Preservation of context, provenance, and
assumptions
• Enhanced fusion and collaboration
• Enhanced support for polymorphism, inheritance,
reification, and inference
History: OGC
• The Open Geospatial Consortium is the epitome
of successful standards bodies
• In 15 years, it has promulgated over 30
specifications
• The GIS industry has adopted and embraced
these standards
• The success of OGC standards has ushered in a
new era of enhanced GIS interoperability,
multidisciplinary activity, and coherent National
Spatial Infrastructure policy
History: CityGML
CityGML is a newly-adopted OGC standard for 3D
mapping of urban terrain
– It adds semantics to Geographic Markup Language
(GML) so that computers (and people) can understand
what structures are made of
– It properly models interior and subterranean spaces
– It properly models the land-structure interface
– It has virtual surfaces so that the volumes of negative
spaces can be computed
– It integrates GIS (maps) with CAD (blueprints)
History: ARML
• Augmented Reality Markup Language is a newlyadopted OGC specification for adding semantics to
computer vision
• For several years, computer vision frameworks such as
OpenCV and Kinect have allowed devices as small as
smart phones to recognize real-world objects in
camera views from COLLADA files (from any angle)
• ARML offers XML semantics to allow applications to
comprehend and act appropriately upon what CV sees
• ARML also offers a JavaScript event-driven model so
that behaviors are easily encoded (or changed in real
time) in thin client applications
Limitations (Until Now)
• CityGML is extremely verbose
– XML can only explain through containment
– Describing a city space means big schemas and big XML instances
• Even small CityGML data sets require large downloads and large
CPU
• The old API did not decouple ARML from COLLADA, so an
implementer had to tackle the whole CV stack
• There was never a way to combine information about what an
object looks like on the outside (COLLADA) with what it is made out
of (CityGML). Implementers had to tie behaviors directly to the
appearance of an object, not the substance of an object
• ARML allows us to divide and conquer: the CityGML folks can focus
on tactical exploitation, the CV industry improves recognition
What Can You Do with CityGML?
•
•
•
•
•
Civil Engineering and Urban Planning
Resilience planning, disaster planning, emergency response
Supply chain, supply chain vulnerability, logistics
MOUT, targeting, damage assessment
Life-after-people longitudinal studies and policy
– Chernobyl, Namie-Dachi
• Back-fill building interiors and subterranean features onto existing
3D urban maps
• Structural failure forensics (I-35 bridge collapse example)
• Role-specific IAPs, SOPs
• Auto-generated AARs
• Dynamic COIs
• eDiscovery (in the legal sense)
CityGML as a Real-World Index
• Its utility is not just in telling you what a structural
component is made of!
• CityGML turns structural components into an index
– Who designed, manufactured, delivered, installed, maintained,
inspected, repaired it? Where are they now and how do I
contact them?
– What equipment, skills or materials are necessary to interact
with it? Where and how do I get them?
– Integration with CIA World Fact Book, World Development
Indicators, Human Geography publications
• CityGML lets analysts locate and quantify anything related
to the structures in an area of interest with a single spatial
query
CityGML as an Index
Supports Personalized Real-Time Plans
• Incident Action Plans with real names,
organization charts, communication plans,
procedures, full FEMA Incident Command System
document generation
• Full Joint Field Office, Unified Incident Command
assignment and staffing
• Full NRF support (NIMS, NIPP, HSIP)
• Full NIEM integration
• ARML adds real-time, computerized vision
prompting via heads-up display or smartphone
Demo
We want to demonstrate the use of CityGML as an index. To do that,
we will perform the following tasks:
1. Embed a link in the CityGML document to an external RDF
resource, using XLINK.
2. Search for and retrieve that link in the CityGML document, using
XQuery.
3. Query the external RDF resource, using SPARQL.
4. Parse the JSON result set, using Javascript.
5. Render the result, using OpenLayers.
6. Get more dynamic, using the Parliament semantic database and
Python.
There are step-by-step instructions at blog.rev-mac.com/node/8
Acronyms
•
•
•
•
•
•
•
•
•
•
•
•
AAR – After Action Report
AOI – Area of Interest
COI – Community of Interest
FGDC – Federal Geographic Data Committee
FSAM – Federal Segment Architecture Methodology
HSIP – Homeland Security Infrastructure Program
IAP – Incident Action Plan
MOUT – Military Operations in Urban Terrain
NIEM – National Information Exchange Model
NIMS – National Incident Managementsystem
NIPP – National Infrastructure Protection Plan
NRF – National Response Framework

similar documents