Remote sensing
Data acquisition
Image processing
  Colour composites
  Geometric corrections
  Radiometric corrections
  Contrast enhancement
  Filtering
  Classification
  Visual interpretation
  Post-classification
  Indices
  Principal Component Analysis
  Combination of images
  Multitemporal analysis
  Fusion of images with different resolution
  Geospatial maps
  Combination of images and other data: DEM and DTM
Radar
GIS
Combination of images

Multi-temporal analysis
 
…then the images change over time

One of the main advantages of satellite remote sensing is to be able to provide observations of the Earth at regular intervals. It thus becomes possible to compare a landscape in different seasons or follow longer-term developments. More than twenty years of library data from satellite readings of the Earth are currently available.

One of the simplest solutions for analysing data recorded on different dates is to display them simultaneously in side-by-side windows. Some image-processing and geographical information system software packages allow one to synchronise the two windows so that a movement or zoom in one of the windows will apply to the other window automatically. It is sometimes also possible to draw in a window (for example, the spread of a new housing development) and have the drawing appear in the two synchronised views.

 

Another possibility for visualising multi-temporal data relies on the additive synthesis of colours by the eye. The human eye interprets the superimposition of the primary colours (red, green, and blue) as new colours covering all the colours of the rainbow. If we superimpose an image acquired on one date and displayed in red on an image acquired on another date and displayed in green, the resulting image will contain colours ranging from black to yellow, with all the hues of green and red in between.


These colours can be associated with the type and intensity of change involved. In the example, we see how images corresponding to two different dates are combined to a multitemporal image wherein the image of date 1 is represented in red and the image of date 2 in green.The areas that were light on date 1 and have become dark appear bright red. The areas that were dark and have become light appear green, and the areas that have remained light appear yellow (red + green = yellow). It is theoretically possible to combine images taken on three different dates by including the blue component as well, but interpreting such images becomes very complex.

Of course, to be integrated into such a multi-temporal image, each image must first be corrected geometrically with regard to a common reference.

One interesting application of multi-temporal processing is to calculate the vegetation index, as explained earlier, for two different dates. Merging the vegetation index images for two different dates will reveal seasonal changes, but also more radical changes, such as deforestation, etc.
1984
1994
In this example, a multitemporal image was created on the basis of Landsat TM 1984 (in red) and 1994 (in green) images of Santa Cruz, Bolivia. The deforestation (in green) that occurred during this period is clearly visible.
Courtesy of NASA - Goddard Space Flight Center/Scientific Visualization Studio.