 |
|
 |
 |
| Adjustment and calibration of satellite images |
|
 |
Before interpreting the downloaded image (raw data) an adjustment of images is needed in order to provide an image as representative as possible of the surface of the Earth. |
As water vapour and other greenhouse gases in the atmosphere absorb and spread electromagnetic radiation, the radiation will be retarded and therefore part of the radiation will not reach the sensors of the satellite. The radiometric correction of measured values (radiometric value) is also important in order to be able to compare several sets of data and consequently provide transaction data.
Each image element may contain "noise" due to among other things:
- Inaccuracy in the sensor system
- Atmospheric conditions giving attenuation and spreading of reflected rays
- Unequal reflection because of variation in the terrain
- Unequal reflection because of the fact that the sun hits the curved surface of the Earth at an unequal angle.
In addition a geometric correction is necessary (geometric straightening) as the recording process makes the image somewhat distorted in relation to the corrected geographic co-ordinates. The distortion is mainly due to the fact that the Earth rotates from west to east when the image is being constructed. As explained above, the image is built up of lines of pixels. As the Earth rotates eastwards, each line in the image will start at a point further west than the previous one.
The geometric precision is among other things influenced by:
- The Earth rotating during registration
- The curving of the Earth
- Variation of altitudes in the terrain
- The satellites are not stable in the orbit
- Lack of stability in the registration mechanism of the satellites
- The projection system of the sensors (central projection)
- Atmospheric conditions
A common procedure for geometric straightening is to identify some checkpoints of the image and specify the geographic co-ordinates of these. Sophisticated image processing programmes have a built-in conversion function where one set of conversion coefficients is calculated from the given checkpoints, and the image becomes automatically corrected and adjusted to a chosen map projection. Geometric correction is necessary in order to be able to use data together with maps or other geographic data. |
| |
It is important to emphasize that there is always a need to check the information from the satellite against what is the reality on the Earth (ground truth). In practice this takes place by having control measurements taken on the Earth in selected localities at the same time as the satellite scans the image.
In connection with surveys of vegetation one may harvest the biomass, weigh it and also record which plants are in the different fields or strips of land. This information is correlated with the satellite data of the same localities. The correlation expresses the connection between what the satellite observes and the biomass/use of land. By repeated measurements during a whole growing season, it is possible to get a survey of the season's total production of biomass, not only that of the checking points, but also of larger areas where the satellite makes observations.
Afterwards an image-processing programme may be used in order to manipulate and interpret the satellite image further and adjust the ground information into the image.
The illustration to the right shows the process from the moment the sensor records the information until the image is ready for analysis (NRS). |
 |
|
|
|
| |
|
|
 |  |
|
|
|
 |
 |
 |