Use of Hyperspectral for Environmental Monitoring of Waste Disposal Areas

1
Use of Hyperspectralfor Environmental
Monitoring of Waste Disposal Areas
Case Study
After Jason Hamel Chester F. Carlson Center for
Imaging Science (CIS) at Rochester Institute of
Technology
2
What are Landfills?
Case Study

• Very common waste management technique
• Toxic wastes are not separated from the
  environment
• A water resistant clay cap is placed over the
  landfill to slow the spread of chemicals

3
Clay Caps
Case Study
4
Clay Cap Technology
Case Study

• Caps are designed to last 40 years
• Need to be replaced with new technology that
  actually deals with the waste
• Meanwhile, chemicals have time to leach into the
  environment

5
Why Look at Landfills?
Case Study

• Currently, possible dangerous sites are manually
  sampled and processed in a lab
• This can be time and money consuming for larger
  sites or a large number of sites
• Chemicals are often dangerous even at very low
  concentrations
• Remote sensing with new hyperspectral detectors
  may provide and economic alternative

6
Example of Expected Imagery
Case Study
Hyperspectral AVIRIS scene with 224 bands DOE
Savannah River Site (SRS)
7
Research Objective
Case Study

• Low concentrations make it very difficult to
  directly detect a chemicals spectral signature
• Determine if new hyperspectral sensors collect
  enough information to identify materials
• Determine the detectability of specific secondary
  spectral effects of leachates (e.g.)
• Vegetation health
• Soil water moisture
• Determine if atmospheric correction is necessary

8
Vegetation Spectra
Case Study

• Two varied inputs
• Chlorophyll concentration (mm/cm2)
• Equivalent water thickness (cm)
• Generated spectra
• Healthy leaf (high chlorophyll and water)
• Stressed leaf (low chlorophyll and water)

9
Vegetation Spectra
Case Study
Reflectance Spectra of Vegetation Green
Healthy Red Stressed
10
Soil Spectra
Case Study

• Ground measurements were taken with spectrometer
  as soil dried
• Moisture in soil was not measured while spectra
  was taken
• Relative labels given to various soil spectra
• Wet Soil
• Moist Soil
• Dry Soil

11
Soil Spectra
Case Study

• Reflectance
• Spectra of Soil
• Brown Dry
• Orange Moist
• Black Wet

12
Reflectance Data Set
Case Study

• The 5 basic vegetation and soil spectra are mixed
  by
• This creates 10 additional mixed spectra
• 15 spectra in final data set

Rmixed 50R1 50R2
where R1 and R2 are 2 basic spectra
13
Atmosphere and Detector Effects
Case Study

• Light reflecting off material propagates through
  atmosphere
• Detector measures the radiance reaching the
  detector at various narrow wavelength regions
  called channels
• Detector electronics record input signal in
  digital counts (DC)

14
AVIRIS Basic DC Spectra
Case Study
15
Realistic Data Set
Case Study

• All detectors measure noise as well as signal
• Standard gaussian noise with standard deviation
  of 1 added to DC spectra (not representative
  AVIRIS noise value)
• Noisy sensor radiance determined
• Noisy reflectance spectra calculated by removing
  atmosphere effects

16
Noisy Basic Reflectance Spectra
Case Study
17
Classification
Case Study

• 6 classification algorithms used
• Linear Spectral Unmixing (ENVI)
• Orthogonal Subspace Projection (Coded)
• Spectral Angle Mapper (ENVI)
• Minimum Distance (ENVI)
• Binary Encoding (ENVI)
• Spectral Signature Matching (Coded)
• The 5 basic vegetation and soil spectra were used
  as endmembers
• Reflectance endmembers converted to DC before
  classifying DC spectra

18
Classification Algorithms
Case Study

• Linear Spectral Unmixing (LSU)
• Generates maps of the fraction of each endmember
  in a pixel
• Orthogonal Subspace Projection (OSP)
• Suppresses background signatures and generates
  fraction maps like the LSU algorithm
• Spectral Angle Mapper (SAM)
• Treats a spectrum like a vector Finds angle
  between spectra
• Minimum Distance (MD)
• A simple Gaussian Maximum Likelihood algorithm
  that does not use class probabilities
• Binary Encoding (BE) and Spectral Signature
  Matching (SSM)
• Bit compare simple binary codes calculated from
  spectra

19
Classification Results
Case Study

• The SAM, MD, BE, and SSM algorithms were not
  designed to classify mixed pixels
• Accuracy is the correct identification of one of
  the fractions in a pixel

Percent Accuracy Classifier Ground Sensor
DC Retrieved
Reflectance with Atmosphere Reflectance
SAM 66.67 40.00
66.67 MD 66.67
80.00 66.67 BE
86.67 66.67 86.67
SSM 93.33 80.00
93.33
20
Conclusions
Case Study

• Atmosphere degrades performance of most of the
  classification algorithms studied
• Removal of the atmosphere is recommended
• The LSU and OSP fraction maps are more useful
• Provide very accurate material identification
  without a large spectral library
• Detects not just the material, but the amount of
  material in a given pixel