xrspatial.multispectral.bai#

xrspatial.multispectral.bai(red_agg: DataArray, nir_agg: DataArray, name='bai')[source]#

Computes Burn Area Index (BAI).

BAI measures the spectral distance of each pixel to a charcoal reflectance point (red=0.1, NIR=0.06). Higher values indicate recently burned areas. Unlike NBR, BAI works on the raw reflectance values rather than a normalized difference.

Input bands should be in reflectance units (0-1 range). If your data is in DN or scaled integers, divide by the appropriate scale factor first.

Parameters:

  • red_agg (xr.DataArray) – 2D array of red band reflectance data (0-1 range).

  • nir_agg (xr.DataArray) – 2D array of near-infrared band reflectance data (0-1 range).

  • name (str, default='bai') – Name of output DataArray.

  • Alternatively

  • first (a single xr.Dataset may be passed as the)

  • Dataset (argument with keyword arguments mapping band names to)

  • example:: (variables. For) – bai(ds, red=’B4’, nir=’B8’)

Returns:

bai_agg – 2D array of bai values. Higher values indicate burned areas. All other input attributes are preserved.

Return type:

xr.DataArray of same type as inputs

References

  • Chuvieco, E., Martin, M.P. and Palacios, A., 2002. Assessment of different spectral conditions for the detection of burned areas with Landsat TM data. International Journal of Remote Sensing, 23(1), pp.71-85.

Examples

>>> import numpy as np
>>> import xarray as xr
>>> from xrspatial.multispectral import bai
>>> red = xr.DataArray(np.array([[0.1, 0.2], [0.3, 0.05]]))
>>> nir = xr.DataArray(np.array([[0.06, 0.3], [0.4, 0.02]]))
>>> bai(red, nir).values  # pixel (0,0) is at charcoal point
array([[       inf, 0.01686341, 0.00858369, 1111.1111  ]],
      dtype=float32)
On this page
Show Source