Midterm 1, Deck pt 2 Flashcards
Data Collection to Raster Analysis
accuracy
degree to which measurement is correct
how much you trust the data
precision
repeatability of the measurement and how small of a scale you are measuring (mm more precise than cm)
2 types of data collection
primary vs secondary
Primary data collection
a direct measurement, you collect yourself
secondary data collection
something collected in advance or using pre-existing data
always check metadata to make sure its trustworthy
T/F: secondary data collection is something we can control
True, we determine what our reolution, precision, accuracy is
examples of primary data collection
surveying land, GPS measurements (using satellites), taking air photos yourself, photogrammetry
examples of secondary data collection
scanning existing vector and raster data, DEMs, gazeteers, digitizing, heads up
Digitizing
converting geographic features of a map into digital format
digitizing tables time-consuming and impractical
digitizing always converts ____ pixels into ______ data
raster, vector
“Heads up” digitizing
digitize scanned maps/documents directly from a computer screen
4 steps of “heads up”
source data, georeference base data, digitize, edit
Georeferencing base data includes ____ or ____
R or R
rectify or register
rectify
rearrange locations to correspond to a specific reference system (coordinates)
map-world
register
rearrangement of locations in one data set to correspond to same locations in another data set
map-map
Linear transformation
everything moves by same amount, distortion is same across image
differential transformation
rubber sheeting
inconsistent stretch in image in different spots, Defining ground control points plotted as polynomial models
Ground control points should be …
easily identifiable, precise, discreet, well distributed, temporally consistent, and ideal for crosshairs
T/F: it is better to use more GCPs than a higher polynomial
True, it reduces error and keeps it less complex
just remeber to keep them well distributed or you’ll only have one detailed section
tie vs tic components in a GCP
Ground control (reference coordinates) vs Map locations (source coordinates)
you “tie” the new points down
Topology
relations used to validate the geometry of points, lines, and polygons
qualitative over quantitative (yes/no)
topographical relations
4
connectivity, adjacency, orientation, containment
connectivity
are two points connected
adjacency
are x and y next to each other
orientation
can we travel in a given direction
containment
is x within y (or vice versa)
arc-node topology
table defining each node that make up a line (arc)
arc-poly topology
table defining arcs that make up each polygon
what are the rows and columns for attribute data tables
records and fields respectively
all records have same fields, one record per observation/entity
4 data types of fields
characters, integers, floats, BLOBs
characters
text or numbers formatted as text “strings”
Integers
numbers without decimals
floats
numbers with decimals
BLOBs
Binary Large OBjects
attachments that are not characters, integers, or floats
references, photos, media, etc
Measurement scale of attribute data
acronym
NOIR
T/F: Raster attribute data tables are detailed
False, you can only have basic raster tables
What is the feature attribute table in regards to table joins
the target table, stores spatial infromation
dbf files
flat files
one large file with all the data in it
pros/cons of flat files
Pro: convenient for simple, ltd data sets
cons: bad for complex data sets, data redundancy (repeats of fields), and data consistency (different data types/spelling/formatting)
Primary key
attribute in the target table that can uniquely identify a record
relations
joining of tables
why are joined tables better than a flat file
they reduce the likelihood of error
what are the 4 cardinalities
target:join
1:1, 1:many, many:1, Many:Many
1:1
one record in target table relates to one record in join table
for every house there is one owner
1:M
one target record relates to multiple join records
for every subdivision there are multiple houses
M:1
multiple target records relate to one join record
multiple houses are in one subdivision
M:M
any combintion of the other 3 relations
outer join
keep all records, keep all inputs in a joined table
inner join
keep only matching records, leaves no empty entries
Structured Query Language
SQL
language to talk about data, provides syntax
Query
Question posed to database in form of a SELECT statement
SQL used for a query: records, table, field
SELECT, FROM, WHERE
Comparison operators
find matches and thresholds, look only for what you need, symbols
<, >, =, not equal
Logical operators
Boolean membership functions
AND, OR, NOT, XOR
FROM is always the ____ ____ table
feature attribute
AND does what to the result sets and includes what in a venn diagram
narrows results, and is the middle only
OR does what to result sets and includes what in a venn diagram
widens results and includes all of the venn diagram
NOT does what to the results and looks like what in a venn diagram
increases precision and includes only one side with no middle
XOR does what to results and looks like what in a venn diagram
satisfies either x or y but not both, includes both sides of venn diagram without the middle
Spatial Joins
match records from a join layer to a target layer
.shp files
spatial joins by distance
attributes of nearest join feature added to target layer along with distance
always 1:1 - you can only have one closest thing
spatial joins by CONTAINS
polygon as target, points as join, attribute table would list points in each polygon record
Spatial join by IS_WITHIN
points as target, polygon as join, for each point record it would list the polygon it is within
is there a change to output layer geometry with spatial joins
No
how are overlays different from spatial joins
geometry of the output changes
UNION overlay
polygon overlay, keeps all features from both datasets (OR)
INTERSECT overlay
Input can be anything but INTERSECT feature must be polygon, features common to all inputs kept, (AND)
IDENTITY overlay
input anything, IDENTITY feature must be polygon, keeps all inputs and changes attributes that intersect identity feature
T/F: dimensionality and extent of output matches input always
True, for all three overlays
Minimum Mappable Unit
arbitrarily decided, smallest entity you want to be represented, a threshold
ELIMINATE tool
slivers < MMU are merged with largest neighbour or largest boundary
non-overlay
CLIP tool
Changes extents of the inputs to fit an area of interest (AOI)
non-overlay
SPLIT tool
split a vector feature into set of contiguous smaller features
non-overlay
DISSOLVE tool
change geometry, turn smaller features into larger ones
(merge municipalities into counties, or counties into states)
non-overlay
Buffering
creates polygons around pts, lines, plygns based on a buffer distance from those features
Raster analysis includes ____ surfaces and ____ gradients
continuous, concentration
continuous surfaces and examples
things that exist everywhere
elevation, temperature, land cover
concentration gradients and examples
phenomena that extends from a point, varying as it distances
smoke, crime stats, pollination, flooding
Logical operators
Boolean conditions, selecting based on criteria
Arithmetic operators
+, -, *, /, tan, sin
“overlay” operators
logical/arithmetic operators on multiple datasets, stacking raster layers together based on pixels
needs column-row coincidence
what is column-row coincidence
when pixels line up, boundaries line up with boundaries perfectly
T/F: in overlay operations, output pixels will all have data even if not all layers have data for a pixel
False. If anything has no data, then that pixel will have no data in the output
Geometric operators
projection, resampling (change in spatial resolution), warping (rubber-sheeting)
map algebra
uses a raster calculator of operators
remember input is always same form as input
3 Scopes of Raster analysis
local, focal, zonal
local scope
operations performed on a cell-by-cell basis, no influence from neighbouring pixels, calculate through layers not across
True or False, yes or no, 0 or 1
Focal scope
moving window
have a focus (kernel) and take some stat and place it in your center pixel of the focus, then move over and do it over and over again
will the output be smaller or larger than the input in a focal raster analysis
smaller, because you lose the outside layer(s) of pixels
Low-pass filter is a ____ filter
smoothing
what does a low-pass filter do
allows low frequency variation to remain, removes high variation and turns it into gradation instead of hard breaks
Low-pass filters take the ____
average, and place it in the center of the kernel
Median filters do what
eliminate extreme values, and remove striping
median filters take the ____
median and put it in the center
high-pass filters
highlight variation, are a sharpening filter
two types of high-pass filters
laplace and sobel
laplace filters
accentuate differences in values between neighbours, by accentuating center pixel, depressing cardinal directions, and diagonals as 0
raises the contrast
sobel filters
look for vertical edges/boundaries between pixels or horizontal edges
zones in raster analysis are …
groups of cells with the same value (code) and gaps are represented by no data cells
can be contiguous or not
zonal raster analysis
finds some stat from each zone and applies it to every corresponding zone cell
