• Home
• Environments
  • Marine
  • Terrestrial
  • Atmospheric
• Interface
  • Text
  • Graphic
• Resources
  • What is remote sensing
  • Fundamentals
  • Time and cost
  • Technical info
• Help
• Contacts

BRG GPEM The University of Queensland

• Main page
• Pre-processing
  • Introduction
  • Image acquisition
  • When to correct
  • Correction types
  • Atmospheric correction
  • Relative atmospheric correction
  • Relative atmospheric correction
  • Physics-based correction
  • Bi-directional reflectance correction
  • Bi-directional reflectance correction
  • Geometric corrections
  • When to correct
  • Masking
• Field data collection
  • Introduction
  • Groundcover sampling
  • Field data sources
  • Photo transects
  • DGPS surveys
  • Driving surveys
  • Vegetation structure
  • Using photographs
  • Visual estimates
  • Spectral libraries
  • Example spectra
  • Optical properties
  • Control points
  • Features mapping
• Mapping approaches
  • Introduction
  • Manual delineation
  • Object based analysis
  • Supervised classification
  • Unsupervised classification
  • Inverse modelling
  • Example
  • Regression analysis
• Change and trends
  • Introduction
  • About
  • Image requirements
  • Mapping change
  • Trend detection
  • Example time series
  • Landsat long term
  • Landsat short term
  • MODIS NDVI time series
  • NDVI trends
  • Example change maps
  • Canopy cover
  • River width
  • Land use
  • Foliage cover
  • Vegetation colour
• Delivery
  • Introduction
  • Post classification
  • Refinements and smoothing operations
  • Stratification and post-classification sorting
  • Error assessment
  • Sampling design
  • Sampling unit size
  • Number of samples
  • Sampling scheme
  • Selecting a reference data set
  • Reference data example
  • Derivation of error or contingency matrix
  • Category level accuracy
  • Map level accuracy
  • Quantitative image products
  • Interpreting mapping accuracy
  • Interpreting mapping accuracy
  • Interpreting mapping accuracy
  • Interpreting mapping accuracy
  • Final map creation

For relative atmospheric correction it is assumed that the haze component present in an image is additive (evident in x-axis offset in histogram):

1. Locate area in image with 0 reflectance
2. Record DN_dark for 0 reflectance feature in each spectral band
3. Subtract DN_{dark-band_n} from all pixels in band n

References:

Jensen, JR (2005). Introductory digital image processing: a remote sensing perspective. Third edition, Prentice Hall: 316p.