We need to analyse remote sensing imagery in order to select the specific information from a satellite image that is necessary to interpret it. Which information is regarded relevant is defined by its purpose: for climatology, other priorities are set than for urban geography, for example. The procedure is similar for each purpose: first of all, the image data have to be processed to eliminate atmospheric disturbance, distortion or sensor failures. Afterwards, analysis and interpretation is carried out. Finally, all results have to be checked for plausibility.
After preprocessing, different kinds of analysis can be carried out.
|One example is the visual image interpretation. Here, the image contents are perceived by the observer and put into context. Visual image interpretation can lead to a variety of direct and indirect insights but satellite imagery contains much more information which cannot or only vaguely be observed by humans.|
|For complex data set interpretation, computer-aided analysis is an advantage. This technique will be explained using the example of image classification and change detection.|
|The term classification is usually interpreted as the aggregation of similar pixels to a major class. The reflection values characterizing objects are used to measure similarity - meaning the so-called spectral fingerprint.|
|Change detection deals with changing landscapes. Images of an area across different points in time are analysed with the help of digital image analysis so that changes in the course of time are visualized (e.g. glacial recession or desertification).|