4. 1Challenge
4
Challenges associated with Environmental Management:
• vast areas requiring active management
• often requires intensive field studies
• time consuming and expensive
5. 1Introduction
5
• Hyperspectral Remote Sensing is an airborne survey of light reflections from the
Earth’s surface, which can be used to identify environmental features and
pollutants.
• Compared to aerial photography – instead of 3 bands (red, blue and green) you
now have 360 band available for analysis.
• This method is very new in South
Africa and is a more accurate
methodology to identifying certain
features on the Earth’s surface.
6. 1Data collection
6
Aerial survey:
• Simultaneous acquisition of 360 bands covering the full Visible Near
InfraRed (VNIR) and Short Wave-InfraRed (SWIR) portions of the
electromagnetic spectrum (390-2450nm)
• Imagery was acquired at 1.5m spatial resolution
Field survey:
• A field survey was conducted after each flight to derive spectral signatures
for the various target features
• A calibrated, handheld ASD Spectrometer was used to record spectral
signatures and corresponding GPS locations in the field
7. 1Data Processing
7
• Data is stored in a “DATA CUBE” which has a layer per wave length band.
This data cube forms a solid environmental baseline with a time stamp
8. 1Hyperspectral Signature
8
• Signatures enable the
differentiation with plant
species, mineral compositions
and qualities.
• Using HYPERSPECTRAL
SIGNATURES, we can extract
the location of certain elements
(such as minerals,
hydrocarbons, alien vegetation
etc.) from our DATA CUBE
• It is known that many features existing on the Earth’s surface reflect certain
amounts of energy at certain wavelengths, this is called a HYPERSPECTRAL
SIGNATURE. During previous academic research worldwide, a
comprehensive dictionary of HYPERSPECTRAL SIGNATURES has been
compiled.
10. 2Application of technique
10
• Glencore undertook a pilot study in April 2011.
• The aim of the study was to utilise this remote sensing technique to identify
the following:
• Hydrocarbon spillages (Coal/ oil/ diesel)
• Alien vegetation including Black Wattle, Eucalyptus and Pampas Grass
species
• Fine coals/ discard/ coal slurry covered by vegetation
• Test vegetation health on rehabilitated areas
11. 2Application of technique
11
• The pilot study proved successful in identifying black wattle, eucalyptus, the
probability of hydrocarbon spillages and using NDVI to determine vegetation
health
• Based on the success of the pilot study, Glencore conducted a full survey of
the previous Xstrata Mining areas in November 2012
14. 3Identification of alien vegetation species
14
Black Wattle (Acacia mearnsii) and the Blue Gum
(Eucalyptus globulus) are trees indigenous to
Australia and are considered widespread invasive
alien trees in South Africa.
Each of these species has a unique hyperspectral
signature and this has been used for measuring
leaf area indices (LAI’s) in the forestry industry to
determine the overall health of plantations of Black
Wattle and Blue Gum
The same methods can be applied to alien species
populations in areas where there is a need to
control alien vegetation infestations
20. 4Rehabilitation Monitoring - Vegetation Health
20
• Normalised Difference Vegetation Index (NDVI) was used as a measure of
relative health.
• NDVI is mathematically defined as follows:
• NDVI = (NIR – Red) / (NIR + Red)
Relative NDVI
• The cooler colours (blue and green)
denote low vegetation activity or poor
vegetation health,
• The warmer colours (red and orange)
represent good vegetation health.
36. 6Conclusion
36
• Alien Vegetation
• Large tracts easily identifiable using remote sensing
• Small solitary species (pampas) not successfully isolated yet
• Able to indentify developing (future) tracts of alien vegetation
• Vegetation Health on rehabilitated sites
• Remote sensing can isolate problem areas on these sites
• Creates focus areas for further investigation and potential maintenance
• Hyperspectral remote sensing can be successfully used to monitor
environmental conditions on Coal Mining sites as well as reduce costs by using
this information to focus efforts for maximum benefits
