Blob labels

We can create an additional field showing blob labels by setting the labels argument to either same or individual when instantiating the Model object. same will label regions where blobs are present to 1 and the background to zero. individual will give individual integers as labels for each blob. The resulting filed will be stored as blob_labels in the xarray dataset. This option can be useful for creating a training dataset for supervised machine learning algorithms.

The borders of the blob labels are defined with the label_border argument. The label regions are located where density >= label_border * amplitude.

Let’s take a look at an example: Let’s say we want to calculate the individual blob labels of some Gaussian blobs:

from blobmodel import Model
import numpy as np
import matplotlib.pyplot as plt


bm = Model(
    Nx=100,
    Ny=100,
    Lx=20,
    Ly=20,
    dt=0.1,
    T=20,
    periodic_y=True,
    blob_shape="gauss",
    num_blobs=10,
    t_drain=1e10,
    labels="individual",
    label_border=0.75
)

ds = bm.make_realization(speed_up=True, error=1e-2)

ds['n'].isel(t=-1).plot()
plt.figure()
ds['blob_labels'].isel(t=-1).plot()

plt.show()