Download Rasterization

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Computer simulation wikipedia , lookup

Theoretical computer science wikipedia , lookup

Neuroinformatics wikipedia , lookup

Multidimensional empirical mode decomposition wikipedia , lookup

Data analysis wikipedia , lookup

K-nearest neighbors algorithm wikipedia , lookup

Pattern recognition wikipedia , lookup

Geographic information system wikipedia , lookup

Data assimilation wikipedia , lookup

Corecursion wikipedia , lookup

Transcript
Raster Data Models: Data Compression

Why?
– Save disk space by reducing information
content
– Methods
•
•
•
•
Run-length codes
Raster chain codes
Block codes
Quadtrees
Raster Data Compression Models:
Run-length Encoding
Reduces data volume on a row-by-row basis by indicating string
lengths for various values
From An Introduction to Geographic Information Systems, Heywood et al. (2002)
Raster Data Compression Models

Quadtrees: Recursively divide an area into
quadrants until all the quadrants (at all
levels) are homogeneous
1
1
2
2
1
1
2
2
3
3
2
2
3
3
3
3
NW
1
NE
2
SW
3
SE
2
2
3
3
Comparison between Raster and Vector Data Structure
Real World Features
Raster
Vector
•
•
•
A Raster model tells what occurs everywhere, while a vector model tells where everything occurs.
Which Model to Choose: Raster or Vector?
It depends on:
spatial data characteristics:
Vector model: linear features such as rivers, roads, political boundaries
Raster model: temperature, elevation
the software available:
less restrictive now than a few years before. Most software handles
both vector and raster data
the training of the key individuals
historical precedent
many big GIS projects are continuous works for many years, and many
new works depend on what’s have been done and what’s been collected
Advantages and Disadvantages
Advantages of Vector Model:
Allows accurate measurement of areas, distance, and geographic representation
less demanding for data storage
Allows topology, thus facilitating error detection in the database
Disadvantages of Vector Model:
Data acquisition and input are time consuming
Not suitable for continuous Geographic phenomena
Computationally intensive for some spatial analysis
Advantages and Disadvantages
Advantages of Raster Model:
Simple and straight forward data structure
Suitable for continuous geographic phenomena
Computationally efficient for spatial analysis
Disadvantages of Raster Model:
Data redundancy for homogeneous areas, thus large data volume
Limited accuracy of location, thus corresponding areas, distance
Vectorization and Rasterization
Data conversion are required in many applications due to
multiple layers of data may come from different sources
that do not meet the application needs.
Vectorization: conversion of raster data to vector format
Rasterization: conversion of vector data to raster format
Vectorization
Vector data are often preferred because of the following:
• smaller file size, and thus easier handling by computers
• for flexible to adjust to different scale
• more efficient to handle topology
e.g. common edge can be easily define base on
left and right side polygons
Vector data format is often preferred in most GIS software and projects
Vectorization
Points & polys - relatively simple
points: if cell=value, then a vector point is created at cell
centroid with attribute=value
polygons: polygon with attribute=value is created for all
adjoining cells=value; poly boundary follows exterior of cells
Lines - more complex
must somehow determine:
• start/end/intersection points (nodes) for lines
• shape points along lines (vertices)
• topological relationships
Vectorization
Starting point: Raster representation of vector world
Vectoriation Steps
Step 1
Line thinning
Line thinning: thin multicell linear features to one cell width
Vectoriation Steps
Step 2
Line extraction
Line extraction:
• determine locations of line begin/end intersections
-- generate nodes
• determine line shape points
-- generate vertices
Vectoriation Steps
Step 3: assign x-y coordinate to each nodes and vertex
raster origin
Y-axis
(x0,y0)
x3=X0+2*cellsize
y2=y0-2*cellsize
Real world origin
(0,0)
(2,2)
X-axis
Vectoriation Steps
Step 4: construct topology
Determine how to connect
the nodes and vertices
Information Loss
Lost information
Rasterization
Important Issues:
cell size
cell values
Simpler compare to
vectorization.
Rasterization
Cell size:
Smaller cell size, better representation of vector data, more
data – fine spatial resolution
Larger cell size, less data, but more information loss!
-- coarse spatial resolution
How to determine cell value
• cell center (centroid)
• majority weighting
• weighted values based on priority/importance
Rasterization
value at centroid assigned to the cell
simple, but can over-represent small area values
Rasterization
value covering the majority of area assigned to cell
simple; “fairer” representation than centroid approach
Rasterization
analyst assigns priority weights based on importance of different values
“most important” value present is assigned to cell
assures presence of crucial geographic phenomena