Data Models (Raster) Reading Assignment: Bolstad, Chapter 2.

Post on 18-Dec-2015

220 views 1 download

Transcript of Data Models (Raster) Reading Assignment: Bolstad, Chapter 2.

Data Models (Raster)Reading Assignment: Bolstad, Chapter 2

Rasters (Digital Photos)

Zoomed In

Pixelation

Raster = Matrix of Pixels

Rasters...• Similar to a digital photograph• Natural for scanned or remotely sensed data.• Can Represent:

– continuous surfaces (e.g., topography)– Or categorical (land cover types)

• Can take much more space• Rasters provide uniform (same cell size)

coverage for an area

Raster Data Model

Uses grid cells of a given dimension to represent the value or attribute of a real world entity or phenomenon.

• In a raster layer, the cells are arranged in rows and columns.

• All cells in a raster layer are the same size.

• Cell Size = Spatial Resolution: the dimension of the area covered on the ground and represented by a single cell (e.g., 10m).

• The value may be a measurement or a code.

• Cell values are numeric: can be either positive or negative, integer, or floating point.

Name ArcGISAttributes

ArcGIS GRIDS* Geodatabase

Bit 1 bit

Chew 2 bit

Nibble 4 bit

Unsigned byte Unsigned 8 bit

Signed byte Signed 8 bit

Unsigned short Unsigned 16 bit

Signed short Short Integer Signed 16 bit Short Integer

Unsigned Integer Unsigned 32 bit

Signed Integer Signed 32bit Long integer

Long Long Integer

Float Float Floating-point 32 bit Single-precision floating point

Double Double Double-precision floating-point

String Text Text

Date Date Date

* ArcGIS documentation indicates the GRID values are always stored as 32-bit valuesSee: http://www.esri.com/news/arcuser/1002/files/table_2.pdf, http://webhelp.esri.com/arcgisdesktop/9.3/index.cfm?TopicName=Bit_depth_capacity_for_raster_dataset_cells ,http://webhelp.esri.com/arcgisdesktop/9.3/index.cfm?topicname=Technical_specifications_for_raster_dataset_formats

ArcGIS Data Types

Name Number of Bits

Number of Bytes

Minimum Value Maximum Value

Number of Values

Sig. Digits

Bit 1 1/8 0 1 2 (21) <1

Chew 2 ¼ 0 3 4 (22) <1

Nibble 4 ½ 0 15 16 (24)

Unsigned Byte 8 1 0 255 256 (28) >2

Signed Byte (aka chars)

8 1 -128 127 256 (28) >2

Unsigned Short 16 2 -32768 32767 65536 (216 or 64k) >4

Signed Short 16 2 0 65535 65536 (216 or 64k) >4

Unsigned Integer (Int)

32 4 0 4,294,967,295 4,294,967,296(232 or 4 Gig)

>9

Signed Integer 32 4 -2,147,483,648 2,147,483,647 4,294,967,295 >9

Long(always signed)

64 8 A big negative number

A big positive number

264 >19

Float(always signed)

32 4 ~10-40 ~1040 232 ~7

Double(always signed)

64 8 ~10-300 ~10300 264 ~15

See: http://en.wikipedia.org/wiki/Integer_overflow, http://steve.hollasch.net/cgindex/coding/ieeefloat.html

Computer-Based Numeric Data Types

Statewide GIS layer (GRID format) of the “working landscape” (areas managed for ag/timber/forage, urban and residential areas, public and private ownership, and reserves )

Images:

True Color Composite (multi-band raster data set; 3 raster layers; 1 each for RGB)

Spectral Reflectance

Raster Coordinate Data

• Coordinates of a corner location are stored (origin).

– This and cell dimension are used to calculate location of other cells.

Column

Xcell= Xlower-left + column * cell size

Xcell4 = 100,000 + (4 * 10) = 100,040

0 1 2 4 53

?

Raster File Formats– All standard image formats (JPG, TIFF, GIF)– Imagine– Mr. Sid– ESRI GRID

Conversion between data models

Vector Raster

Conversion between data models

Choosing an appropriate cell size is not always simple. You must balance your application's need for spatial resolution with practical requirements for quick display, processing time, and storage.

Problems associated with conversion:

– Loss of Detail– Loss of Accuracy– Stair Stepping (raster to vector)– Changes to the original data

Choosing between data models

Often depends on

– Type of entity or phenomena represented. – (discrete or continuous)

– Available storage. – Expected types of analysis. – Expertise of human operators.– Level of accuracy desired.

“Raster is faster but vector is corrector”

Choosing between data modelsRaster is useful when:

– Working with continuous data types– Good for large area analyses – Good for surface analysis– Mathematical modeling– Spatial detail isn’t important

Vector is useful when:– Working with discrete data types – Good for small study areas – Spatial detail is important (When “close enough” isn’t really

good enough). – When topology is needed for the analysis