DESCRIPTION

i.hyper.continuum applies continuum removal to hyperspectral imagery imported as 3D raster maps (raster_3d) by i.hyper.import.

The module performs continuum removal, a spectral processing technique that normalizes reflectance spectra to emphasize absorption features. Continuum removal divides the original spectrum by a continuum line (or envelope) that connects the spectral peaks, effectively removing the overall spectral shape and highlighting absorption bands.

i.hyper.continuum is part of the i.hyper module family designed for hyperspectral data import, processing, and analysis in GRASS. It is essential for mineral identification, vegetation analysis, and any application requiring detailed analysis of spectral absorption features.

The module allows users to specify a wavelength range for processing, producing a spectral subset of the original hyperspectral image. By default, it processes the entire spectral range of the input data. Bands outside the specified range can optionally be kept unchanged in the output.

Three methods are available for calculating the continuum:

• convex_hull : Uses convex hull algorithm to find the upper envelope of the spectrum (default, recommended for most applications)
• linear : Simple linear continuum connecting the first and last points in the wavelength range
• polynomial : Polynomial fitting to the spectral envelope (order 2-5)

The convex hull method is most widely used in remote sensing as it effectively captures the spectral continuum without making assumptions about the shape. It identifies the upper envelope by constructing the convex hull of the spectral points and extracting the upper boundary.

Continuum-removed reflectance values typically range from 0 to 1, where 1 represents points on the continuum (spectral peaks) and values less than 1 represent absorption features. Deeper absorptions produce values closer to 0.

NOTES

The module expects input data to be a 3D raster map created by i.hyper.import or any 3D raster with wavelength metadata stored in band-level metadata following the i.hyper standard format: wavelength, FWHM, valid, and unit.

When wavelength metadata is not found, the module will fail with an error message. It is essential to use properly imported hyperspectral data with wavelength information.

The wavelength range parameters (min_wavelength and max_wavelength) define both the spectral region for continuum calculation and the extent of the output 3D raster. If not specified, the module uses the minimum and maximum wavelengths available in the input data.

The -n flag restricts processing to only bands marked as valid (valid=1) in the metadata. This is useful for excluding bands with quality issues, such as those affected by atmospheric absorption (water vapor bands around 1400 nm and 1900 nm) or sensor artifacts.

The -k flag preserves original reflectance values for bands outside the specified wavelength range. Without this flag, all bands in the output maintain the structure of the input, but only bands within the range undergo continuum removal. This is useful when you want to analyze a specific absorption feature while maintaining the full spectral cube structure.

The -i flag provides information mode that displays band information and processing parameters without creating the output raster. This helps verify which bands will be processed before running the full analysis.

For the polynomial method, the order parameter controls the degree of the polynomial fit. Lower orders (2-3) produce smoother continua, while higher orders (4-5) can follow more complex spectral shapes but may overfit noise. The convex hull method is generally preferred as it adapts naturally to the spectral shape without requiring parameter tuning.

Continuum removal is particularly effective for:

• Mineral identification: Absorption features at specific wavelengths indicate different minerals (e.g., clay minerals at 2200 nm, iron oxides at 900 nm)
• Vegetation analysis: Chlorophyll absorption around 680 nm, water content features at 1450 nm and 1950 nm
• Comparison across scenes: Normalizing for illumination and atmospheric effects
• Feature depth analysis: Quantifying absorption strength for compositional mapping

The output 3D raster maintains the same spatial resolution and extent as the input, with continuum-removed reflectance values replacing the original reflectance in the processed wavelength range.

EXAMPLES

# Apply continuum removal to entire spectral range using convex hull
i.hyper.continuum input=prisma \
                  output=prisma_cr

# The output contains continuum-removed reflectance across all wavelengths

# Focus on clay mineral absorption feature (2100-2400 nm)
i.hyper.continuum input=enmap \
                  output=enmap_clay_cr \
                  min_wavelength=2100 \
                  max_wavelength=2400

# Extract specific band for visualization
# Band depth at 2200 nm indicates clay mineral presence
r3.cross.rast input=enmap_clay_cr output=clay_2200nm@2200

# Apply linear continuum removal for faster processing
i.hyper.continuum input=tanager \
                  output=tanager_cr_linear \
                  method=linear \
                  min_wavelength=400 \
                  max_wavelength=2500

# Polynomial continuum with 3rd order fit
i.hyper.continuum input=prisma \
                  output=prisma_cr_poly3 \
                  method=polynomial \
                  polynomial_order=3 \
                  min_wavelength=400 \
                  max_wavelength=2500

# Process only iron oxide absorption region (800-1000 nm)
# Keep other wavelengths unchanged
i.hyper.continuum input=enmap \
                  output=enmap_fe_cr \
                  min_wavelength=800 \
                  max_wavelength=1000 \
                  -k

# The -k flag preserves original values outside 800-1000 nm

# Preview processing parameters without creating output
i.hyper.continuum input=prisma \
                  output=test \
                  min_wavelength=2000 \
                  max_wavelength=2400 \
                  -i

# Info mode displays which bands will be processed

# Use only valid bands for processing
i.hyper.continuum input=enmap \
                  output=enmap_cr_valid \
                  min_wavelength=400 \
                  max_wavelength=2500 \
                  -n

# Excludes bands with valid=0 (atmospheric absorption, sensor issues)

# Complete workflow: mineral mapping with continuum removal
# Step 1: Import hyperspectral data
i.hyper.import input=/data/PRISMA.he5 \
               product=prisma \
               output=prisma

# Step 2: Apply continuum removal to Al-OH region (clay minerals)
i.hyper.continuum input=prisma \
                  output=prisma_aloh_cr \
                  min_wavelength=2000 \
                  max_wavelength=2400

# Step 3: Extract band at 2200 nm (main clay absorption)
r3.cross.rast input=prisma_aloh_cr output=clay_depth@2200

# Step 4: Calculate absorption depth (1 - continuum_removed_reflectance)
r.mapcalc "absorption_depth = 1.0 - clay_depth"

# Step 5: Classify clay mineral abundance
r.recode input=absorption_depth output=clay_class << EOF
0.0:0.1:1:low
0.1:0.2:2:moderate
0.2:1.0:3:high
EOF

# Step 6: Visualize results
r.colors map=absorption_depth color=viridis
d.rast map=absorption_depth

# Vegetation water content analysis
# Step 1: Apply continuum removal to water absorption region
i.hyper.continuum input=enmap \
                  output=enmap_water_cr \
                  min_wavelength=1400 \
                  max_wavelength=2000

# Step 2: Extract bands at water absorption centers
r3.cross.rast input=enmap_water_cr output=water_1450@1450
r3.cross.rast input=enmap_water_cr output=water_1950@1950

# Step 3: Calculate water absorption index
r.mapcalc "water_index = (1.0 - water_1450) + (1.0 - water_1950)"

# Higher values indicate stronger water absorption (higher water content)

# Compare different continuum methods
# Convex hull (default)
i.hyper.continuum input=prisma output=prisma_cr_hull \
                  min_wavelength=2000 max_wavelength=2400

# Linear
i.hyper.continuum input=prisma output=prisma_cr_linear \
                  method=linear \
                  min_wavelength=2000 max_wavelength=2400

# Polynomial order 2
i.hyper.continuum input=prisma output=prisma_cr_poly2 \
                  method=polynomial polynomial_order=2 \
                  min_wavelength=2000 max_wavelength=2400

# Extract same band from each for comparison
r3.cross.rast input=prisma_cr_hull output=hull_2200@2200
r3.cross.rast input=prisma_cr_linear output=linear_2200@2200
r3.cross.rast input=prisma_cr_poly2 output=poly2_2200@2200

SEE ALSO

REFERENCES

• Clark, R. N., & Roush, T. L. (1984). Reflectance spectroscopy: Quantitative analysis techniques for remote sensing applications. Journal of Geophysical Research: Solid Earth, 89(B7), 6329-6340.
• Kokaly, R. F., & Clark, R. N. (1999). Spectroscopic determination of leaf biochemistry using band-depth analysis of absorption features and stepwise multiple linear regression. Remote Sensing of Environment, 67(3), 267-287.
• van der Meer, F. (2004). Analysis of spectral absorption features in hyperspectral imagery. International Journal of Applied Earth Observation and Geoinformation, 5(1), 55-68.
• Kruse, F. A., et al. (1993). The spectral image processing system (SIPS)—interactive visualization and analysis of imaging spectrometer data. Remote Sensing of Environment, 44(2-3), 145-163.

AUTHORS

Created for the i.hyper module family

Based on work by Alen Mangafić and Tomaž Žagar, Geodetic Institute of Slovenia