This preview shows page 1. Sign up to view the full content.
Unformatted text preview: Raster data: resolution and Raster data: resolution and remote sensing Raster data Raster data
• tessellation data model: a matrix of cells (pixels) of some regular tessellation, each of which carries some information (value) • raster = rectangular tessellation (squares) • why squares?
– looks like a computer monitor – borrows from computer science theory of image processing – looks like grid coordinate systems base maps continuous surfaces datadefined polygons (thematic maps) Raster data Raster data
• values of different characteristics at the same location are stored in different layers or bands
– each spatial unit (a pixel) in a raster layer in GIS has one variable (a color) – in vector GIS, each spatial feature (a point, line, polygon) can have many characteristics • layers record different themes (forest, drainage, highways) • bands record different wavelengths (red, infrared, UV) Characteristics of raster data Characteristics of raster data
• values can be categorical (land use) or numerical (elevation)
– often value attribute tables (VATs) match numerical data to categorical data • values can apply to the center of, or the whole cell Characteristics of raster data Characteristics of raster data
• raster data are stored as a long list of values
80, 74, 62, 45, 45, 34, 39, 56, 80, 74, 74... – header part of the file indicates length and width – each cell represents an equal portion of the mapped area • • • areal extent of layer / number of pixels = spatial resolution
the smaller (“higher”) the spatial resolution, the more detail and smaller the object that can be seen how many pixels are required to give a spatial resolution of 1 square foot if our area is 100 square feet? What is the spatial resolution of a raster layer with extent 100 feet by 100 feet if it is covered by:
– 40,000 pixels? – 100 pixels? Spatial resolution Spatial resolution
• parallel to data scale of vector data Note how the measurement of the area of the feature varies with resolution • more not always better – your results will be only as accurate as your dataset with the lowest resolution QDJ 5: question 1 QDJ 5: question 1
if a raster layer covers 100 square kilometers in 10,000 pixels, what is the area of the smallest feature that can be resolved? Give an example of a realworld object of this size. Map scale and spatial resolution Map scale and spatial resolution
same data scale (resolution), different map scale same map scale, different data scale (resolution) Other types of resolution Other types of resolution
• spatial resolution is only one type of raster resolution • temporal resolution: frequency of data collection
– how often does a satellite in orbit revisit the same place? – how frequently are aerial surveys taken? – how many images are taken by a surveillance camera every second? Other types of resolution Other types of resolution
• radiometric resolution: the smallest difference in energy that can be detected by the sensor
– how many different colors (or shades of gray) can be captured? – also known as the bit depth (or, in photography, the contrast)
does the sensor record values from 0 to 99? from 0 to 255? Other types of resolution Other types of resolution
• spectral resolution: the electromagnetic wavelengths to which a sensor is sensitive • vital in image classification: – identification of features or classes in raster layers • many satellites and other sensors can pick up information in many layers at once digression: remote sensing basics Remote sensing Remote sensing
• Collecting, storing, and extracting environmental information using devices not in contact with what’s being studied • Human vision is a form of remotesensing…limited by:
– Inexact storage and recall – Our nonvertical perspective – Small portion of the electromagnetic spectrum that our eyes can discern Electromagnetic radiation (ER) Electromagnetic radiation (ER)
• All objects with temperature above absolute zero emit ER over a broad range wavelengths • The range of possible ER λ s is the EM spectrum • remote sensing instruments operate in all regions of the ER spectrum (exc. Radio, X, and gamma ray) ER: absorption and reflection ER: absorption and reflection by ground objects
• Selective absorption and reflection due to molecular composition of object – A cyan object absorbs almost all red light but reflects radiation in the blue and green wavelengths • We can distinguish objects based on their characteristic patterns of reflectance: their spectral signatures Spectral signatures Spectral signatures
• objects reflect different visible colors, but also different amounts of other types of ER • land use classification based on analysis of several bands of radiation • LANDSAT can detect seven bands of data Same place, different wavelengths QDJ 5, question 2 QDJ 5, question 2
• you have the following spectral signatures and data from a satellite • classify the following pixels (wavelength, value): – (1.0 µ m, 0.25; 2.5 µ m, 0.20) – (1.0 µ m, 0.35; 2.5 µ m, 0.18) – (1.0 µ m, 0.72; 2.5 µ m, 0.02) ...
View Full Document
This note was uploaded on 11/30/2009 for the course GEOG 3561 taught by Professor Robedsall during the Spring '09 term at Minnesota.
- Spring '09