Computes the spectral gradient over all bands of an image (or the first band if the image is Array typed) by computing the per-pixel difference between the spectral erosion and dilation with a given structuring kernel and distance metric. See: Plaza, Antonio, et al. 'Spatial/spectral endmember extraction by multidimensional morphological operations.' IEEE transactions on geoscience and remote sensing 40.9 (2002): 2025-2041.
The input image.
|String, default: "sam"|
The spectral distance metric to use. One of 'sam' (spectral angle mapper), 'sid' (spectral information divergence), 'sed' (squared euclidean distance), or 'emd' (earth movers distance).
|Kernel, default: null|
Connectedness kernel. Defaults to a square of radius 1 (8-way connected).
|Boolean, default: false|
If true, distances are computed from the mean of all pixels under the kernel instead of the kernel's center pixel.