csquares a new simple, - PowerPoint PPT Presentation

Loading...

PPT – csquares a new simple, PowerPoint presentation | free to download - id: 143648-MThmO



Loading


The Adobe Flash plugin is needed to view this content

Get the plugin now

View by Category
About This Presentation
Title:

csquares a new simple,

Description:

Characteristics of metadata, and metadata spatial searches ... (structured dataset descriptions) - as ... used for distributed spatial searching, 1995 onwards ... – PowerPoint PPT presentation

Number of Views:40
Avg rating:3.0/5.0
Slides: 40
Provided by: ree103
Learn more at: http://www.marine.csiro.au
Category:

less

Write a Comment
User Comments (0)
Transcript and Presenter's Notes

Title: csquares a new simple,


1
c-squares - a new simple, XML friendly, display/
query/ exchange format for representing spatial
data extents at the metadata level System
concept and development by Tony Rees Divisional
Data Centre CSIRO Marine Research, Australia
2
Topics to be covered ...
  • Characteristics of metadata, and metadata
    spatial searches
  • Problems with bounding rectangles as
    representations of dataset extents
  • The c-squares concept
  • c-squares in practice
  • Future possibilities

3
Metadata, and spatial searching of metadata
records
4
The Metadata concept ...
(Data level)
Data Store 1
Data Store 2
offline digital data
databases / data warehouses
offline nondigital data
5
some example Metadatabases (Data Directories) ...
many others -- 100 lt 1000? ...
  • Metadata records exist independently of the
    datasets they describe, may not necessarily have
    on-line connection to the actual data --- i.e.,
    they act as surrogates for the data
  • Spatial searching (where implemented) typically
    by bounding rectangles (N,S,W,E limits) or
    sometimes defined regions (R1 yes/no, R2 yes/no,
    etc.)

6
current first pass representation of spatial
data coverage is by bounding coordinates -
example
ltmetadatagt lttitlegtFranklin Voyage FR 10/87 CTD
Datalt/titlegt ltcustodianOrggtCSIRO Marine
Researchlt/custodianOrggt (etc. etc.)
ltboundingBoxgt ltnorthBoundingCoordgt-9.0lt/no
rthBoundingCoordgt ltsouthBoundingCoordgt-19.
0lt/southBoundingCoordgt ltwestBoundingCoordgt
117.0lt/westBoundingCoordgt
lteastBoundingCoordgt145.8lt/eastBoundingCoordgt
lt/boundingBoxgt (etc. etc.)
  • concept introduced in FGDC draft metadata
    standard, 1994
  • used for distributed spatial searching, 1995
    onwards
  • still the primary tool for conducting metadata
    spatial searches integral to ISO 19115 draft
    metadata standard, 2002
  • polygons are also enterable, but seldom used for
    searching owing to the arithmetic overhead
    involved

7
Bounding coordinates - pluses and minuses
  • Pluses ...
  • Metadata elements are concise
  • User-entry is simple
  • Spatial searching is simple arithmetic operation
    (looks for overlap between a search rectangle
    and available data rectangles)
  • Useful as a first pass -- rapidly filters out
    many datasets not close to the region of interest
  • Minuses
  • A rectangular shape does not correspond to the
    actual shape of many datasets
  • Data distribution may be aligned along other
    than N-S or E-W axes
  • Data distribution may be patchy or incomplete
    within the designated boundary
  • Corollary Apparent hits never 100 reliable
    (unless the data are always rectangular, e.g.
    mapsheets)

8
Some real-world examples (other agencies data)
...
9
our agencys data (marine surveys) - examples ...
NB, bounding rectangle searches result in many
false or misleading hits, since large portions of
the dataset rectangles contain no data -
particularly where surveys wrap around a feature
or land area, or are oriented obliquely with
respect to N-S, or E-W directions.
10
Germ of c-squares concept ... from Ken Walkers
Bioinformatics search interface, Museum Victoria
(Australia)
  • state divided into 0.5 x 0.5 º squares (numbered
    as per relevant mapsheets)
  • search interface has direct connection to base
    data (gt100,000 point data records)
  • each base data record is tagged with its
    relevant mapsheet number, so spatial searching is
    by simple numeric/text match (no arithmetic
    required)
  • user can request list of hits (species) from one
    or multiple search squares (e.g. blue hatched
    examples)

700 km
11
modifications which would be required for use
with metadata ...
  • multiple square ids could be stored in single
    metadata record (harvested from base data) -
    removes requirement to access the base data to
    answer search queries
  • numbering system should be expanded to become
    globally applicable
  • geographic scale (size of squares) should be
    variable up or down to suit variety of user needs
  • metadata records become storage vehicles for
    dataset footprints (simple spatial objects)

700 km
12
The c-squares concept c-squares Concise
Spatial Query and Representation System
13
c-squares principle
  • c-squares string holds IDs of all the tiles
    (e.g. 1 x 1, 0.5 x 0.5 degree squares) which are
    intersected by the dataset spatial extent
    (footprint)

actual survey location - Franklin cruise 10/87
data footprint using bounding rectangle
14
c-squares numbering system
  • each square is numbered according to a globally
    applicable system based on recursive divisions of
    WMO (World Meteorological organisation) 10-degree
    squares, e.g.
  • 10 degree
    square 3414 ( WMO number)
  • 5 degree square 34142
  • 1 degree square 3414227
  • 0.5 degree square
    34142274
  • 0.1 degree square
    3414227466
  • (etc.)
  • strings of codes represent an individual dataset
    extent, e.g.
  • 301349731114683111478311147931114883111
    489311149931121223112123
  • 311213131121323112134311214131121423112
    143311221731122183112219
  • 311222631122353112350311235131123523112
    353311236031123613112362
  • 311236331123703112371311238031123813112
    390311310031131013113102
  • 311310331131043113205311320631132073113
    216311321731132283113238
  • 3113239
  • encodes the extent
  • shown in the example

15
WMO 10-degree squares notation (part)
(Available via the web in NODC, 1998 World
Ocean Database 1998 Documentation)
16
WMO 10-degree squares notation principle
NE sector (1xxx)
7800
NW sector (7xxx)
7000
7017
1017
3000
3017
5000
5017
SE sector (3xxx)
SW sector (5xxx)
3800
5800
17
nomenclature for 5-degree squares - e.g. in SE
sector
  • follows Blue Pages (1996) extension of WMO
    numbering, using 4 quadrants (1, 2, 3, 4) for
    5-degree squares - e.g. within 10-degree square
    3414 ...

140
145
150
-40
WMO 10-degree square 3414 (grey) 5-degree
square 34142 (light blue)
1
2
-45
4
3
3414
-50
(1 is always closest to global origin, 4 is
always furthest away. For full specification
refer c-squares website)
18
nomenclature for 1-degree squares - e.g. in SE
sector
  • follows Blue Pages (1996) extension of WMO
    numbering, using 4 quadrants (1, 2, 3, 4) for
    5-degree squares, plus 2 digits 00-99 for
    1-degree squares - e.g. within 10-degree square
    3414 ...

140
145
150
-40
100
101
102
103
104
205
206
207
208
209
WMO 10-degree square 3414 (grey) 5-degree
square 34142 (light blue) 1-degree square
3414227 (green)
110
111
112
113
114
215
216
217
218
219
120
121
122
123
124
225
226
227
228
229
238
130
131
132
133
134
235
236
237
239
140
141
142
143
144
245
246
247
248
249
-45
350
351
352
353
354
455
456
457
458
449
360
361
362
363
364
465
466
467
468
469
370
371
372
373
374
475
476
477
478
479
380
381
382
383
384
485
486
487
488
489
3414
390
391
392
393
394
495
496
497
498
499
-50
(100 is always closest to global origin, 499 is
always furthest away. For full specification
refer c-squares website)
19
Codes have straightforward relationship with
lats/longs, mapsheets, etc. ...
e.g. 3414227 (1-degree square with origin at
42 º S, 147 º E)
additional degrees E
1407 147 additional
degrees S 402 42
5-degree quadrant, i.e. 1 2

3 4 tens of degrees E (i.e.,
140) tens of degrees S (i.e., 40) global
sector (1NE, 3SE, 5SW, 7NW)
70 km
20
quad tree -type approach used where numerous
adjacent squares are occupied
example 3212 can be used instead of
specifying every 1-degree square within 10 degree
square 3212. This leads to corresponding data
reduction, e.g. Australia (at 1-degree
resolution) can be described in 343 squares
rather than 800
21
Example database-level implementation of
c-squares for metadata records (e.g. at 1 degree
resolution)
(etc.)
22
Options for c-squares data entry ...
  • automated conversion of lat/long data to
    c-squares (ignoring multiple hits)
  • automated conversion of GIS polygon data to
    c-squares extents
  • clickable map interface for region(s) of
    immediate interest
  • manual entry, with reference to marked-up
    mapsheet/s
  • on-line lat/long - to - c-square converter
  • custom digitising system (graphics tablet data
    input or similar)

clickable map interface (generalised example)
mapsheet marked with 0.5 degree squares - for
manual entry
23
Process invoked for web mapping (1)
c-squares strings can be transformed into
coordinate pairs (centre point of squares) and
square size, by an appropriate function and then
sent to Xerox PARC Map Viewer or similar, e.g.
24
Process invoked for web mapping (2)
c-squares strings can be sent directly to the CMR
c-squares mapper (accessible via the web), e.g.
25
Further examples (CMR oceanographic/biological
data - 0.5 x 0.5 deg. squares)
(Base maps are automatically chosen to fit the
data range, or can be selected manually)
26
Mechanism for spatial queries using c-squares
  • c-squares spatial queries simply test whether a
    text string representing the search box (ideally
    one or several c-squares) is matched anywhere in
    the c-squares string
  • example - search square 31132 will match any
    c-squares string which includes 31132 within it,
    e.g.
  • ltcsquaresgt
  • 301349731114683111478311147931114883111
    489311149931121223112123
  • 311213131121323112134311214131121423112
    143311221731122183112219
  • 311222631122353112350311235131123523112
    353311236031123613112362
  • 311236331123703112371311238031123813112
    390311310031131013113102
  • 311310331131043113205311320631132073113
    216311321731132283113238
  • 3113239
  • lt/csquaresgt
  • (NB, this is a simple text search and involves no
    arithmetic - cf. querying of bounding rectangles,
    polygons, or more complex spatial objects)
  • hierarchical naming system for c-squares means
    that finer resolution squares are automatically
    picked up in any coarser resolution search

27
Implementable as a simple click on a square
interface, e.g.
28
system does the search - checks for c-squares
match if available (provides reliable matches),
otherwise uses overlapping rectangles test
(possible match) ...
29
produces ...
(etc.)
30
Viewing the full metadata record produces ...
with clickable link to show dataset extent using
c-squares
(etc.)
31
Base maps for displayed data can be changed at
will by the user, e.g.
(numerous other maps available, sample only shown)
32
c-squares strings are suitable for inclusion as a
new XML metadata element, for example ...
ltmetadatagt lttitlegtFranklin Voyage FR 10/87 CTD
Datalt/titlegt ltcustodianOrggtCSIRO Marine
Researchlt/custodianOrggt (etc. etc.)
ltboundingBoxgt ltnorthBoundingCoordgt-9.0lt/no
rthBoundingCoordgt ltsouthBoundingCoordgt-19.
0lt/southBoundingCoordgt ltwestBoundingCoordgt
117.0lt/westBoundingCoordgt
lteastBoundingCoordgt145.8lt/eastBoundingCoordgt
lt/boundingBoxgt ltcsquaresgt311149923112390131
1148933112380331123804311238113111488
231123812311237133111478431123704311
237013111478131114792311147913112361
4311146843112363331123613311146723112
360231123631311236223112360131123524
311235233112350431123521311235123112
35223112353231123531lt/csquaresgt lt/metadata
gt
33
Actual size of c-squares, e.g. compared to U.K.
WMO Square 7500
7500 1000 x 600 km
10 x 10 deg.
75001 500 x 300 km
5 x 5 deg.
1 x 1 deg.
7500123 100 x 60 km
75001234 50 x 30 km
0.5 x 0.5 deg.
0.1 x 0.1 deg.
7500123455 10 x 6 km
(NB, real shape and dimensions vary according
to position on globe)
  • 1 x 1 degree squares is suggested as a possible
    minimum standard of spatial encoding for global
    interoperability of metadata systems (finer
    resolution available to users on as-needs basis)

34
Summary - strengths and weaknesses of c-squares
  • Strengths ...
  • c-squares metadata element is a concise and
    flexible way of encoding a wide variety of
    different spatial objects - including nonlinear
    and incomplete (patchy) coverages
  • automated or manual code entry (and maintenance)
    is possible, and relatively simple
  • spatial searching is simple text string matching
    operation -- no supporting GIS system is required
    ( i.e., zero technological overhead)
  • c-squares mapper utility provides rapid and
    flexible data extent visualisations, and can be
    called from anywhere via the web
  • can be implemented progressively into any
    metadata system as an adjunct to bounding
    coordinates (a search can be configured to work
    with whatever is available)
  • Weaknesses
  • may not be the only numbering convention
    available (Marsden Squares and Maidenhead
    Locators are alternatives to WMO squares, however
    less suitable in this application)
  • c-squares are not uniform shape/size across the
    earths surface (true squares only at the
    equator) some local/national grids do not
    transform easily to lat/long squares
  • may be cumbersome to encode very large, complex
    regions (e.g. Pacific Ocean) by this method -
    works best at continental scales and below.

35
other comments ...
  • c-squares notation is language-independent -
    can be equally used in English, French, Japanese
    also discipline-independent (suitable for
    physical, biological, geological, topographical,
    plus any other data type)
  • downwards-scalability of the c-squares notation
    means that it can be applied to any size region
    (e.g. local level)
  • equally applicable to terrestrial and marine
    data
  • no equivalent in GML notation at this time (GML
    only supports vector data). Even if there were a
    GML equivalent, c-squares would still be
    significantly more concise.

36
c-squares future ...
  • c-squares is being implemented progressively in
    CSIRO Marine Researchs MarLIN metadata system
    (c. 500 records to date, more continuously added)
    and in the CMR CAAB marine species dictionary
    (c. 3000 records). MarLIN c-squares search
    interface is already operational
  • c-squares is freely available for implementation
    in any other agencies metadata systems. Possibly
    small islands of interoperability could be
    created, or system could simply be implemented
    for within-agency use
  • c-squares could be offered to relevant user
    community/national bodies as an optional metadata
    element - possibly as a user-defined extension to
    a recognised metadata standard (e.g. ANZLIC, ISO)
  • current CMR c-squares mapper is already
    accessible for general use. Global and selected
    regional mapping options already available and
    can be developed further. External systems
    already linking to the c-squares mapper include
    OBIS (Ocean Biogeographic Information System,
    USA) and FishBase (ICLARM/FAO), as well as CMRs
    MarLIN and CAAB databases
  • c-squares website (www.marine.csiro.au/csquares/)
    is a focal point for all c-squares related
    materials - including specification, background
    information, sample code, on-line lat/long
    converter, sample c-squares-enabled metadata
    records, and more

37
Potential Implementation across multiple systems
Single or multi catalogue query with c-squares
Single or multi catalogue query with c-squares
metadata query and/or exchange with c-squares
bounding rectangles
catalogue 1
metadata query and/or exchange with bounding
rectangles
catalogue 2
(c-squares enabled - whole or part)
(non c-squares enabled)
Single or multi catalogue query with bounding
rectangles
catalogue 3
38
Acknowledgements/Inspiration ...
  • Ken Walker (Museum Victoria) for showing me his
    Museum Victoria Bioinformatics search interface,
    based on 0.5 degree squares
  • Blue Pages Marine and Coastal Data Directory
    (MCDD) for the notation for subdividing WMO
    squares, also for pointers to software for
    drawing rectangles on GIF images (as used in the
    c-squares mapper) and for point-and-click map
    searching
  • CMR Data Centre staff for useful feedback
  • Miroslaw Ryba (CMR) for programming assistance
    with the c-squares mapper
  • John Hockaday (Geoscience Australia) and Doug
    Nebert (FGDC, USA) for helpful comments on
    prototype versions of the system
  • NOAA GLOBE Project and Martin Dix, CSIRO
    Atmospheric Research for provision of backdrop
    images used in the c-squares mapper.

39
Questions, comments?
About PowerShow.com