Lab 8: Spectral Signature Analysis

Introduction:

This assigned lab involved using satellite imagery in ERDAS to check the spectral reflectance signatures of various surfaces around the Eau Claire area.  The purpose of this is to observe the spectral reflectance patterns on a graph and analyze them.


Methods:

An satellite image of the area surrounding Eau Claire provided by professor Cyril Wilson and LANDSAT was brought into a viewer in ERDAS Imagine.  From a set of twelve different Earth surfaces was listed to be found and have their spectral signatures analyzed in a graph.  The twelve surfaces were:  standing water, moving water, vegetation, riparian vegetation, crops, urban grass, dry soil, moist soil, rock, asphalt highway, airport runway, and a parking lot (concrete surface).

Collecting a spectral signature is a rather uncomplicated process that seems rather simplistic.  It involves digitizing a polygon on the specific area of the surface feature desired.  From here, opening up the raster processing tools allows the opening of the signature editor.  The spectral mean plot of the area analyzed can be seen (Figure 1).  This shows the various reflectance levels of the different bands that make up the various spectral wavelengths of the electromagnetic spectrum.  Band one is the blue band, band two is the green, band three is red, band four is near infrared (NIR), band five is shortwave infrared, and band six is thermal infrared.  The levels of these bands varied based on the type of surface analyzed as the reflectance varies greatly based on the material.

The spectral signature of standing water in Lake Wissota was high in the blue band, and near non-existent in the infrared bands, which is to be expected of water as it absorbs nearly all of the energy at the higher wavelengths.    (Figure 1)

All twelve of the aforementioned features were gathered methodically (Figure 2 and Figure 3).  The features were identified by linking Google Earth to ERDAS.  This use of Google Earth as a key aided greatly in locating features such as dry soil and concrete.

All twelve of the required features were gathered and analyzed to see if they matched what was expected as far as reflectance values.  (Figure 2)

The various spectral signatures can all be viewed in one window for comparison purposes.  The greatest variation among the different features appear to be in the shortwave infrared.  (Figure 3)

Conclusion:

The various surface signatures can be easily gathered using ERDAS and the signature editor tool.  These reflectance signatures at times have error in them due to atmospheric interference or other forms of interference in general.  Though correcting for this allows for proper viewing of spectral signatures and can help determine vegetation quality, surface moisture content, or just what the surface is in general (among many other uses).