The study of land cover using satellite images requires tools that facilitate their interpretation, especially those that can integrate multiple raster, allowing the evaluation of the spectral behavior of the coverages and their temporal changes.
For this type of studies it is common to opt for Plugins, however, a proper use of the QGIS Identify tool can meet our requirements.
Multiband data visualization:
Satellite images are generally combined into multiband images.
In the following example two satellite images, a LandSat5 from 2003 and a LandSat8 from 2014, will be used on a cattle farm.
The Identify tool has the following query and display modes:
Consultation Modalities (identify):
In the dialog box, at the bottom in Mode are the options:
Current layer: consults the selected layer.
From top to bottom (stop at the first): consults the Visible layer that is at the top of the Layers Panel.
From top to bottom: consults all Visible layers from the top to the bottom of the Layers Panel.
Layer Selection: shows a dialog to select which visible layer we want to consult.
Types of deployment:
It is possible to view the data in various ways, at the bottom under View:
- Tree: data in the form of a form.
- Table: data in a table.
- Graph: graph of the data (only raster layers).
With the option Grafo the spectral behaviour of a pixel in the different bands of the satellite image is represented graphically.
Comparing images:
In order to compare satellite images, a radiometric correction that equals the dynamic ranges must be carried out beforehand.
For example, in the figure it is not possible to compare the curves due to the different radiometric resolutions of the images:
To compare you must set the Identify dialog to Mode > Up Down and View > Graphs.
The following figure compares corrected images of a growing area with four scenes covering the period from June to 15 September.
Incorporating Plugins.
We also have the option to use the Value Tool (QGIS 2) and Temporal/Espectral Profile Tool plugins.
The figure shows the spectral curve of the 2003 image together with the fluctuation of the average monthly precipitation from January to August using the Temporal/Espectral Profile Tool.
The complements have greater customization options.
You can make this type of query for Raster obtained from interpolation but previously must be combined in a Multiband Raster.