You are on page 1of 18

Raster Data Model

Raster Data Model


⚫ It represents spatial phenomena that vary continuously over the
space such as elevation, land erosion etc.
⚫ Research on raster data model has concentrated on data
structure and compression.
⚫ A wide variety of data used in GIS are encoded in raster such
as satellite images, scanned images etc.
⚫ For data storage and analysis raster is divided into rows,
columns and cells.
⚫ Cells are also called pixels with images.
⚫ Raster represent points with single cells, line with sequence of
neighboring cells and polygon as contiguous cells.
⚫ Raster data are much easier to manipulate, aggregate and
analyze then vector data.
Elements of the raster data model
1. Cell size
1. The cell size determines the resolution of the raster data
model.
2. A large cell size cannot represent the precise location of
spatial feature where small size increases the data volume.
3. Raster Band
⚫ A raster may have a single band or multiple band
⚫ Single band has only one cell value
⚫ Multiband raster has more than one cell value.
Raster Data Structure
Cell-by-Cell Encoding

Run Length Encoding

Quad Tree
1. Cell-by-Cell Encoding
⚫ In this method data is stored as a matrix, and its cell
values are written into a file by row and column.
⚫ Ideal to store the cell values that change continuously,
e.g., DEM, Satellite images etc.
Run Length Encoding
⚫ This method records cell values by row and by group.
⚫ A group refers to adjacent cells with the same cell value.
Quad Tree
⚫ Instead of working along one row at a time, quad tree uses
recursive decomposition to divide a raster into a hierarchy of
quadrants.
⚫ Uses recursive decomposition to divide a grid into a
hierarchy of quadrants.
⚫ Recursive decomposition refers to a process of continuous
subdivision until every quadrant in a quad tree contains only
one cell value.
⚫ Quadrant having cells with the same value will not be
sub-divided, and it is stored as a leaf node.
⚫ Leaf nodes are coded with the value homogeneous quadrant.
⚫ A quadrant having different cell values will be subdivided
until a quadrant at the finer level contains only one value..
Draw Quadtree
Solution
Data Compression
⚫ It refers to the reduction of data volume, important for
data delivery and Internet mapping.
⚫ Data compression are of two types
1. Lossless compression- it allows original image to be
preciously reconstructed. No information is lost in
lossless compression. RLE and packbits are some
compression techniques.
2. Lossy compression- it cannot fully reconstruct the
original image but can achieve high compression rations.
JPEG uses this method.
Data Conversion
⚫ Rasterization – it is the conversion of vector data to raster data.
Following steps are involved.
1. Set up the raster with a specified cell size and assign all cell values
as zero.
2. Set the values of those cells that correspond to point, lines or
polygon boundaries to 1.
3. Fill the interior of the polygon outline with the value 1.
⚫ Vectorization – it is the conversion of raster data to vector data.
Following steps are involved.
1. Line thinning – raster line is thinned identical to 1 cell width.
2. Line extraction – process of determining where line begins and ends.
3. Topological reconstruction – connecting extracted lines and showing
digitizing error.
Types of Raster Data
1. Satellite Images
⚫ Satellite images represent remotely sensed satellite data.
⚫ The spatial resolution of a satellite image relates to the
ground pixel size.
2. USGS Digital Elevation Models (DEMs)
⚫ DEM consists of an array of uniformly spaced elevation
data.
⚫ It is point based, but can be converted to raster data by
placing each elevation point at the center of a cell.

You might also like