Collecting & Using
Geospatial Data in the Field
Digital Government Collaboration
ISU UCSB v NSF v Census BLS USDA USGS
www.statlab.iastate.edu/dg
Geospatial info used throughout
data collection process
Locate
Site
Planning &
Assignments
Collect &
Check Data
Check &
Analyze Process Data
Data
2
NRI Sample Design
Stratified two-stage probability sample
of land in US and territories
– 300,000 area segments (PSUs ~ ¼ section)
– 800,000 points selected within PSUs
– Repeated measurements on sample units
over time
3
Data Collection via Photo-
interpretation & Field Observations
a 386
1 3 b 330
2 9
6
a 330
b 600
a 330
c 386
7
a 378
b 600
d 412
a 391a
5
a 362
4
8 4
a 330
b 600
PSU Measurements
Areas of polygons
with specific land
uses
– Built-up areas
– Farmsteads
– Water
– Federal land
Lengths of linear
features
– Streams
5
Point Measurements
Land cover / use
classification
Agricultural &
conservation practices
Soil and erosion
factors
Ecological
classifications
6
Updating Digital NRI Locations
• History
Sample segment and
point locations
transferred from
maps and photographs
to 7.5’ quadrangle for
digitizing
• Current goal
Update locations with
accurate & precise
measurements 7
Johnson Co Pilot
Historical maps &
data for target
location
Current digital
coordinate
GPS receiver (PPS)
8
Integrate Handheld Data Collection
Instrument & GPS Receiver
9
Integrated User Interface
Receiver set-up, active
link
User’s position &
digital coordinate
Navigation aids
– Turning direction
– Current position relative
to target
– Current path
Accuracy information
– Satellite configuration
Capture button
– Location metadata 10
Satellite
Configuration
11
Location
Metadata
12
Research Team
Computer-assisted survey data collection
Sarah Nusser, Iowa State U
Infrastructures for adaptive systems
Les Miller, ISU
Digital geographic information resources
Mike Goodchild, U Calif, Santa Barbara
Emerging technologies for geospatial data
Keith Clarke, UCSB
13
Research Vision
To formulate, develop & test an extensible
framework that
– supports flexible & effective use of
distributed & heterogeneous geospatial
information resources
– applies to a broad range of field
environments
– takes full advantage of emerging information
technologies
14
Framework Model
Field User
Infrastructure
Repositories Specially Ad Hoc Information
Prepared for Data Resources
Collection Campaign (via Internet)
15
Research Areas
Mediating infrastructures
Field tools (e.g., conflation, sampling)
Wearable technologies
Statistical survey applications
16
Today’s Architecture
Data
Collection
Environment
Data Samples
Collection
Reposi- Attributes
tories
17
External MEDIATION
Reposi- Tiger
tories Dissect
DOQQs query
Maps
DRGs
Data
Search for
Resources Collection
options
Environment
Prioritize
Data Samples results
Conflation
Collection Sampling
Reposi- Attributes Extract
Photos views /
tories metadata
Geospatial
Process,
transform
data for user
18
Iowa Map
Data Server
Mediator wrapper DRGs
Ortho
- Spawns agents
Photos
to get product
NRI Data
User wrapper Repository
- Sends info PSUs
request to
mediator Photos
- Adds field
computing info Compute Server NRI Data
- Spawns agents to Local
get data Cache
- Manipulates, integrates data
19
- Creates view or data (product)
Field Conflation
Merge & compare geographic
information from various sources
while in the field
– Very limited hardware & software
environments
– Resolve registration problems
– Merge any combination of raster and
vector data
20
22
23
24
Towards a general approach to
conflation
Between all data types (raster and vector)
Dealing with all types of disagreement
(positional, semantic)
Running in very small field devices
Account for different objectives when
combining
– Average of multiple good images
– Pick the best as a reference image
– Remove differences via smoothing
25
Emerging Field Technologies
Beyond personal computing
– Personal computer fully engages or occupies user
– Desktops, laptops, most handhelds
Research & explore alternative technologies
for mobile field data collection
– Propotype development
– Interface research
26
Wearable computing
Wearable computers have the potential to
experience the life of the user in a first-
person sense
The computer is working even when the user
is not giving explicit commands (e.g. health
monitors, communication systems)
The user is doing something
besides interacting with
the computer
27
Context awareness
Context (non-explicit user
input) is gathered through “sensing”
the person's environment, state, task
Use context (e.g., where the person is,
what they are doing) to create systems that
proactively support, anticipate, and
facilitate the person's task
Key spatial context is from GPS
28
User Interface for Augmented
Vision
Augmented: see-through
map plus locator
Viewed reality
29
View Options
30
Version 0 of the UCSB self-
contained wearable computer
Consisting of:
CharmIT™ Developer's Kit with a PC/104 platform
GPS module
head-mounted, see-through visual display
Twiddler2 as text input device
Fanny-pack or vest
Version 2 will be web-enabled
31
CharmIT™ Developer's Kit
CharmIT™ is built on the
PC/104 specification, which has
been an industry standard for
embedded computing for
nearly ten years
hundreds of companies
manufacture a wide variety of
PC/104 hardware
majority of components are
low power and ruggedized
CharmIT™ Developer's Kit is
lower cost (approximately
$2000), low power
(approximately 7 watts with
Jumptec 266) and offers
enough computing power for
most everyday wearable tasks
32
Typical CharmIT™ kit configuration:
Customizable, lightweight
aluminum case
Jumptec Pentium 166 or 266
Mhz core board - includes on-
board 10/100 ethernet
USB and SVGA
PCMCIA board with two slots or
SoundBlaster-compatible sound
card
Power conversion/distribution
board
Two Sony NP-F960 batteries
(approx 5.5 hours runtime
each)
All necessary cables and
connectors
33
Head-mounted displays
Integrated Eyeglassisplays ($5000)
ClipOn Display ($2500)
-evaluation kit comes with a belt-worn, VGA interface
box connected to the display by a 4’ cable
Display format: 640x480, 24-Bit color, 60 Hz refresh rate
Field of View: Approximately 16 degrees horizontal
34
Text input
The Twiddler2 chorded
keyboard is designed for one-
handed input with an array of
12 finger keys and six thumb
keys.
Frequent users can enter text
at close to two-hand touch-
typing speeds.
35
Wearable computer vests
36
Future applications…
Eyetracker is essentially an eye driven cursor or "visual
mouse" that positions the cursor in the direction that the eye
is looking. The eye's ability to rapidly shift gaze in a
changing environment makes it an ideal input device for
computer and video applications.
Nogatech Video Capture Products
USB Digital Video Camera CaptureVision™
Imagine real-time full motion video capture and display
plus high resolution still frame capture on a notebook…
Geospatial Information as
Reference
Settings
– Locating sample outlets & households
– Locating unmarked field points
Questions
– Strategies for using map & GPS
info in planning, navigation, locating
– Role & effectiveness of alternative
geospatial info resources
– Appropriate materials, formats
– Impact of screen size, interface mode
38
NRI Testbed
9 sq mi area in
north Ames
Transitioning from
riparian &
agricultural land to
urban development
3 faux PSUs
39
iPaq w/ wireless connection
40
41
Collecting & Using
Geospatial Data in the Field
Digital Government Collaboration
ISU UCSB v NSF v Census BLS USDA USGS
www.statlab.iastate.edu/dg