Data Pre-Processing - Geographic Information Systems - Lecture 6 | CSS 4200, Study notes of Agricultural engineering

Download Data Pre-Processing - Geographic Information Systems - Lecture 6 | CSS 4200 and more Study notes Agricultural engineering in PDF only on Docsity!

CSS 4200

Geographic Information Systems Lecture 6: –^

Data Pre-processing

Data Acquisition

1. Preprocessing 2. Data Management3. Spatial Analysis and Manipulation

Data Processing

Information Output

Format Conversion

-^

Vector to raster

-^

Raster to vector

-^

Transform formats– interchange with other GIS packages– to or from accepted or

de facto

standards

(DRG, DLG, DIME, TIGER, DXF, etc.)

• to or from more general standards (SDTS) -^

Transform data types and file types withinparticular GIS software program

(.e00, .shp, geodatabase)

Cell size (m) = [(msd, m

2 )/(#grid cells)]

msd = minimum size delineation (from Burrough and McDonnell 1998)

Vector to Raster Conversion

Geometric Transformation

-^

Transform the geometric characteristics of datalayers

-^

Rectification vs. registration– rectification: refers to

correcting

geometric

relationships

• registration: refers to

matching

data layers

• both follow a common conceptual process -^

Usually performed in preprocessing and output

Rectification

-^

Rectification– processing to produce geometric relationships that

are more correct (defining & removing distortions)

• general rectification process

1 select a general model of the geometric relationship between

the data and the desired base 2 calibrate model using homologous points or features (e.g.

Ground Control Points, GCPs) to produce a specific model 3 check specific model using additional (independent)

homologous points or features (Check Points) 4 if acceptable, apply specific model to the entire data set

Registration

-^

Attempt to match the geometry of a data layer toanother

-^

Making a location in one layer correspond theappropriate location in another layer– the other layer may be less correct geometrically but

have preferable characteristics

• registering a layer to a more correct standard will

provide a degree of rectification

-^

Rectified data layers do not necessarily register(unless rectification also performed registration)

Models of Geometry

-^

Interpolative (approximate) Models– approximate the geometric relationship using

relatively simple mathematical expressions that donot attempt to describe the geometric or physicalproperties of the relationship

• often use general polynomials of some selected

degree

• many also use a “finite element” approach that

interpolates between nearest control points

• these models provide basis for procedures known as

“ rubber-sheeting

Models of Geometry

(con’t)

-^

Hybrid Models– include properties of both interpolative and geometric

models

• include simplified geometric properties but do not

attempt to be rigorous in geometric or physicaldescriptions

• accomplished by considering simple mathematical

approaches to scaling, rotation, translation, etc.

• these models remain simple, usually low order

polynomials

• example: four parameter (conformal) transformation

Models of Geometry

-^

General Idea:

) ,

(

) ,

(

Y

X

f

Y

Y

X

f

X

X Y

= ′

= ′

c

bY

aX

X

= ′

(^2).

(^543) ,

450

(^3).

28

(^3). 5

−

−

= ′^

Y

X

X

-^

Example of a General Model

-^

Example of a Specific Model

Generalization and Aggregation

-^

Basic Ideas– often more data or detail than needed to adequately

perform the GIS analysis

• attempt to reduce the amount of data while

maintaining the information needed by the projects

• foolish to create and maintain a spatial database that is too

complicated, detailed, and large

• must be careful if the database is designed for long-term use• applies to attribute as well as graphic aspects of spatial data

Generalization and Aggregation

-^

Generalization (thinning, smoothing)– remove extra points or cells along linear features,

OR

• remove isolated features or excessive detail -^

Aggregation– combine many detailed spatial objects into fewer,

more general spatial objects

• aggregation is also considered a method in Spatial

Analysis and Manipulation

-^

Often need capabilities for merging features (lineremoval, “eliminate, “dissolve”)

Building Topology

-^

Process of relating spatial objects (e.g. points,lines, areas) to each other

-^

Encoding higher order spatial objects from moreprimitive objects

(e.g. from points to lines, polygons)

• Raster topology is systematic and (usually) simple– Vector topology involves filling in tables for each type

of spatial object

-^

Object-oriented analysis continues this idea tohigher levels

-^

Conclude with error detection and editing

Error Detection and Editing

-^

Functions that add, delete, or change values orpositions of spatial objects

-^

Quality or utility of editing functions is importantin judging the quality of a GIS software program

-^

Two standards:– initial data processing

• verification (detection of problems)• editing (correcting or accepting problems)
  • updating