Vector and Raster Representation

Although the Spring process data in the raster and vector format, sometimes it is required to convert data from one format to another because some operation can only by applied in an specific format.

To use a thematic information layer in a geographic analysis, that is, a fields algebra for example, it is required that the IL is represented in the raster format or raster or even a Thematic Image. Therefore it has to execute a Conversion Vector-Raster.

In other cases it is required to make a Conversion Raster-Vector. For example, when it is desirable to change a class or edit lines from a polygon present in an IL that has just a thematic image, resulting from satellite image classification.

Vector and raster representations were designed for different data models, but the Conversion between raster and vector implemented in the Spring depends on the data model being considered.


Vector Representation

An object vector representation is a way to represent the object as close as possible from the real object, trying to define precisely all the positions, lengths and dimensions of the geographic entities. This format is, in general, a result from the maps digitization. It consists, normally, of lists with 2D coordinates limiting thematic regions, or representing networks that can have associated to it a third measure. Examples: vegetation cover map format and terrain elevation map format (contour lines).

The vector representation uses the following basic elements:

  • Points: Geographical entities identified by a single pair of coordinates (x,y). Usually, a point is a symbol related to any geographic entity that can not be represented in its own dimension (area).
  • Arcs: Set of coordinates (x,y) that describes a continuous line in space. It is used to represent entities that have just a length dimension, or as polygon limits.
  • Nodes: The initial and final points in each arc. To these nodes it is associated a topology information (the occurring lines).
  • Polygons: It is related to regions limited by arcs. Each polygon is built from a list of arcs defining it.

These elements are better defined in Vector Data Editing.

In the Spring the categories, from different data models, that can have representations in the vector format are presented in the table below:

Category/Model

Vector Representation

Example

Thematic

Points, Lines and Polygons

Numerical

Samples (contour lines and spot heights spots) and

TIN (triangular grid)

Cadastral

Points, Lines and Polygons

Network

Points and Lines


Raster Representation

The raster or raster format is defined as a set of cells located in sequential coordinates, implemented in a 2D matrix. Each cell, also called image element, matrix element or pixel, is referenced by row and column indexes and has a number representing the type or value of the mapped attribute.

These matrix values can be limited in a certain numerical interval, as an example the 0 to 255 interval for 8 bits images, or numbers associated to classes in the thematic image case, or even real values obtained through mathematical interpolation in numerical modeling analysis.

The satellite image acquisition mechanics, such as the Landsat and Spot, allow that the image are captured in the raster format and stored in the digital format. Lately, in a laboratory, these digital images can be processed and printed in photographic paper. The opposite process happens with analogical photographic images, obtained by aerophotogrammetric methods, that can be discretized using scanners and stored in the digital format.

The raster (or raster) and vector representations are not exactly equivalent for the same data. Normally there is a lost of precision when converting from the vector format to the raster format, because continuous boundaries are discretized according to the output image resolution. This lost can be compensated because the geographic analysis operations are more efficient in the raster domain.

The vector representation is the most adequate to identify objects, individualized in the terrain, where precision is required. The raster representation is more adequate to phenomena and measures that continuously change in space.


The following table presents the different data models, that can present representations in the raster format.

Category/Model

Raster Representation

Example

Thematic

Thematic image

  • A pixel – a point
  • Aligned Pixels – a line
  • Grouped Pixels - polygons

Numerical

Rectangular Grids

  • Real values associated to each matrix point

Image

Monochromatic Image
  • Pixels with gray levels,
Synthetic Image (coded)
  • Pixels associated to the table of colors
Classified Image
  • Pixels Group with the same color


In the Spring only the thematic maps, as an example, soil map, vegetation, geology, etc.., can be represented using either the vector or the raster representation. Considering that the vector is the closest representation of the entities mapped, in order to get the required precision in the raster, mainly in analysis process, the used resolution has to be smaller than the scale precision in a given project. However the smaller the resolution, the larger the disk space required to store the thematic image.


The table below shows a comparison between the advantages and disadvantages about storing raster and vector for thematic maps.


Vector representation
Raster representation
Advantages
  • Map represented in the original resolution
  • Attributes association to graphical elements
  • Topological relationships
  • Adequate for larger scales (1:25.000 and larger)

Problems
  • It does not represent phenomena with continuous variations in the space
  • Simulation and modeling is harder
Advantages
  • Represents variant phenomena in space
  • Simulation and modeling is easier
  • Fast geographic analysis
  • Adequate for small scales (1:50.000 and smaller)

Problems
  • Amount of disk space used
  • Possible lost of resolution and hard to associate attributes


See the procedures to convert the representation in thematic maps.


See also:
Digitalization of Maps in SPRING
Spatial analysis