GIS/GPS Chapter 5

Data model concepts

simplified view fo real world, physical entities/phenomena area approximated by data in GIS with spatial location and extent of physical entities and non-spatial properties, entity is represented by spatial feature/spatial object, essential characteristics are also defined for each entity, objects are abstractions in a spatial database, spatial objects

Spatial objects

objects in a spatial database representing real-world entities with associated attributes

Spatial data

describes location (where), don’t have a million attributes for any type of data, never put everything into 1 single file due to having trouble keeping it organized

Attribute data

specifies characteristics at that location (what, how, much, when, etc)

How do we represent data digitally in GIS?

group into layers based on similar characteristics using a vector or faster model, selecting appropriate data properties for each layer with respect to projection, scale, accuracy, and resolution, database management systems don’t exist in CAD

Vector data model

coverage with ARC/INFO, shapefile with ArcView

Faster data model

GRID/Image with ARC/INFO and ArcView

Spatial data types

continuous, areas, networks, points/lines/areas

Areas

unbounded with land use and rock types, bounded with zoning areas and counties, moving with air masses and animal herds, networks with groves, utilitizes, and streams, points/lines/areas being fixed like wells, streets, or lamps and moving with cars, planes, and any living organism

Nomial

no order, soil types, county names

Ordinal

ordering of things, roads, zip codes, streams

Binary

this/that = red/blue, yes/no, landslide/not landslide

Discrete

only take on certain values, categorized into a classification scheme

Continuous

interval lacks an absolute 0, ratio has a defined 0, income, height, and weight

Goal of relational DBMS

produce map of values by district/neighborhood

Problem with relational DBMS

no district code available in parcel Table

Solution with relational DBMS

join parcel table, containing values with geography table containing location codings, using Block as key field

Secondary/foreign key with relational DBMS

common info in both tables

Field types

default to a shorter integer due to being smaller bytes to use less storage space with different precisions, text, date, BLOBs, object identifers, global identifiers (GUID)

Text

up to 64,000 characters, text strings such as names and descriptions, could be numerical classifications

Date

Mm/dd/yyyy, hh:mm:ss, AM/PM

BLOBs (binary large objects)

annotation and dimensions, images and other multimedia, use custom loader/viewer

Object identifiers

ObjectID field, guarantees a unique ID for each row in the table

Global identifiers (GUID)

registry style string consisting of 36 characters enclosed in curly brackets, uniquely identify a row within a geodatabase

Raster

location is referenced by a grid cell in a rectangular array (matrix), attribute is presented as a single value for that cell, faster than vector, as zoom in more it’s more pixelated, data comes from images from remote sensing, scanned maps, and elevation data from USGS, best for continuous features with elevation, temp, soil type, and land use

Vector

location referenced by x, y coordinates that can be linked to form lines and polygons, attributes referenced through unique ID number to tables, more correct/accurate than raster, data comes from DIME and TIGER files from US Census and DLG from USGS for streams, roads and cansus data, best for features with discrete boundaries like property line, political boundaries, and transportation

Representing data using raster

area covered by grid with equalized cells, location of each cell calculated from origin of grid (2 down and 3 over), pixels are cells, image data/imagery are raster data, attributes are recorded by assigning each cell a single value based on the majority feature in the cell like land use type, easy to do overlays/aalyses by combining and corresponding cell values, yield = rainfall + fertilizer

Raster orientation

angle between true borth and direction defined by raster columns

Class

set of cells with same value (type = sandy soil)

Zone

set of contiguous cells with same value

Neighborhood

set of cells adjacent to a target cell in some systematic manner

Resolution and storage size

say we have an 8-bit image that’s 500 columns by 500 rows, storage size is 500 × 500 × 8 = 2,000,000 bytes, 2,000 kbytes, 2 Mbytes

Tesselations

square grid, rectangular, triangular and hexagonal, triangulated irregular network

Square grid

equal length sides, conceptual simplest, cells can be recursively divided into cells of same shape, 4-connected neighborhood with all neighboring cells being equidistant, 8-connected neighborhood with all neighboring cells not equidistant and center of cells on diagonal is 1.41 unites away

Rectangular

commonly occurs for lat/long when project, data collected at 1 degree by 1 degree will be varying sized rectangles

Triangular and hexagonal

all adjacent cells and points are equidistant

Triangulated irregular network (tin)

vector model used to represent continuous surfaces (elevation) and more later under vector

Example of tessellations

neverending data

What does the size of the square grid divided evenly on a recursive base?

length decreases by half, number of areas increases fourfold, areas decrease by 1/4

What does resampling do with combining the 4 cell values?

storage increases if you save all samples but can save processing costs if some operations don’t need high resolution

How does nominal/binary data save storage?

using maximum block representation with all blacks with same value at any one level in tree can be stored as a single value

Band sequential (BSQ)

each characteristic in separate file, elevation file, temp tile, good for compression, good if focus on one area

Band interleaved by pixel (BIP)

all measurements for a pixel grouped together, good if focus on multiple characteristics of geographical area, bad if you want to remove/add a layer

Band interleaved by line (BIL)

rows follow each other for each characteristics, not good at compressing anything

Raster data structures database representation

raw data may come in BSQ, BIP, BIL but not good for efficient GIS processing, represented as standard database table, joins based on ID as the key field can be sued to relate variables in different tables

Generic raster data model is actually implemented in several different computer file formats

GRID is ESRI’s propietary format for storing and processing raster data only in ESRI, standard industry formats for image formats can be used to display raster data but not for analysis, georeferencing information required to display images with mapped vector data that requires an accompanying world file to provide lacational information, geotiff

Geotiff

single file incorporates both the image and the world information into a single file

Point (node)

0-dimension, single x, y coordinate pair, 0 area, tree, oil well, label location

Line (arc, polyline)

1-dimension, 2+ connected x, y coordiantes, road and stream

Polygon

2-dimension, 4+ ordered and connected x, y coordinates (4 minimum), first and last x, y pairs are the same, encloses an area, census tracts, county, and lakes

Whole polygon (boundary structure)

polygons described by listing coordinates of point in order as you walk around the outside boundary of the polygon, all data stored in 1 file inefficiently store attribute data for polygon in samefile, coordinates/borders for adjacent polygons store 2x may not be same but results in silvers (gaps)/overlaps, all lines are double (except for those outside the periphery), no topological information about polygons, used by 1st computer mapping program, SYMAP in late 60s, adopted by SAS/GRAPH and many business thematic mapping programs

Points and polygons

polygons described by listing ID numbers of points in order as you talk around the outside boundary a second file lists all points and their coordiantes, solves duplicate doordinate/double border problem, lines can be handles similar to polygons, no topological information, used 1st by CALFORM, 2nd generation mapping package from Laboratory for Computer Graphics and Spatial Analysis at harvard in early 70s and good for saving room

Node/arc/polygon topology

topological components which permit relationships between spatial elements to be defined (always going clockwise)

ARC

defines relations between points by specifying which are connected to form arc, defines relationships between arcs by specifying which arc are connected to form routes and networks

Polygon

define polygons (areas) by specifying which arcs comprise their boundary

Left-right

defines from nodes and to nodes that permit, need a topology file separated, left permit and right polygon to be specified

Features in theme (coverage) have unique identifiers

common identifiers with links to coordinate table and attribute table, concepts are those of a relational database and really prerequisite for vector model

Points

elevation points (nodes) chosen based on relief complexity and their 3D location determined

Polygons

elevation points connected to form a set of triangular polygons then represented in a vector structure

Attribute

attribute via relational DBMS

Coverage

multiple physical files (12+) in folder, proprietary

KML/KMZ

keyhole markup language, KMZ is zipped KML file and associated files

OpenStreetMap

crowdsourced GIS data project, XML based file formats, .osm.pbf denotes a OSM protocolbuffer binary format

Shapefile

comprised of several files all of which must be present, additional files with .sbn and .sbx, openly published specs to other vendors can create 2-5 files

Geodatabase

multiple layers saved in single .gdb

Object view

real world is a series of entities loathed in space, object is digital representation of an entity with 3 types of point, line, area objects, some entity is represented at different scaled by different object types, behavior can be associated with objects

Field view

real world has properties which vary continuously over space, raster/vector, value is categorical/integer variable then places with same value can be grouped, floating point data can categorize and treat it as an object

Representing surfaces

surfaces involve 3rd elevation value to x, y horizontal values, complex to represent since there are an infinite number of potential points to model with raster-based digital elevation model and vector based triangulated irregular networks, ¾ alternative digital terrain model approaches available, massed points and breaklines

Raster-based digital elevation model

regular spaced of elevation

Vector based triangulated irregular networks

irregular triangles with elevations at the 3 corners

Vector-base contour lines

lines joining points of equal elevation at specified Harvard

Massed points and breaklines

raw data from which 1 regular/irregularly spaced point elevations, breaklines

Breaklines

point elevations along a line of signifcant change in slope (vallet, floor, ridge, crest)

Digital elevation model

sample array of elevations that are regularly spaced intervals in x and y directions, 2 approaches for determining surface z value of location between sample points and lattice each mesh point represents a value on the surface at the center of the grid approximated by interpolation between adjacent sample points and doesn’t imply an area of constant value, surface grid considers each sample as a square cell with constant surface value