Example - `acorns` clustering in 2D¶

Here is a simple example for running acorns in 2D, e.g. dust continuum data. This isn’t really what acorns was designed for, but it can do it anyway, so here’s how…

The data used for this example can be downloaded here. For reference, this is a 3.2mm dust continuum image taken using the Plateau de Bure Interferometer (now NOEMA). You can find more details about the data in Henshaw et al. (2016b).

Start by loading in the data:

import numpy as np
from acorns import Acorns
import sys
import os
from astropy.io import fits

datadirectory =  './'
datafilename =  datadirectory+'G035_cont.fits'

# Load Continuum data
hdu   = fits.open(datafilename)
header = hdu[0].header
data  = hdu[0].data
hdu.close()
data = np.squeeze(data)

# Noise level of the continuum
rmsnoise = 7.e-5

acorns works by reading in a table. Because we have a 2D image here, in the case of our continuum image, we first of all just want to flatten this:

# Create the acorns table
x = np.arange(np.size(data[0,:]))
y = np.arange(np.size(data[:,0]))
xx,yy = np.meshgrid(x,y)
# Flatten 2D arrays
xx = xx.flatten(order='F')
yy = yy.flatten(order='F')
data = data.flatten(order='F')
# Create an equivlent length noise array
noisearr = np.ones(len(xx))*rmsnoise
# This is the array that will be fed to acorns
dataarr_acorns = np.array([xx,yy,data,noisearr])

Now set up the input parameters:

# Basic information required for clustering
pixel_size = 1.0
min_radius = 1.7 # Ensures 9 pixels defines the smallest structure identified
min_height = 3.0*noisearr[0] # Clusters have to be at least this value above the merge level

# Generate the cluster_criteria - In this instance we are only going to use
# a single linking length - i.e. the distance between two pixels.
cluster_criteria = np.array([min_radius])
# Relax criteria - In the case of continuum data we don't need to relax the
# linking constraints as all pixels are adjacent.
relax = np.array([0])
# Stopping criteria - determines when the algorithm will stop - this is a
# multiple of the rms value
stop = 3.

Now call acorns:

# Call acorns
A = Acorns.process( dataarr_acorns, cluster_criteria, method = "PP", \
                    min_height = min_height, pixel_size = pixel_size, \
                    relax=relax, stop = stop, verbose=True )
A.save_to(datadirectory+'example_2d.acorn')

If verbose=True, then the output in the terminal should look like this:

Beginning analysis...

acorns will look for clusters within 1.66015625% of the data:

Method = PP
Max. Euclidean distance between linked data = 1.48492
Min. number of data points in a cluster = 9
Min. height above the merge level = 0.0
Min. separation between clusters = 2.96985
Primary clustering...

[==================================================>] 100%

Making second pass...

[==================================================>] 100%

Generating forest...

[==================================================>] 100%

acorns took 2.0 seconds for completion.
Primary clustering took 2.0 seconds for completion.

acorns found a total of 48 clusters.

A total of 2176 data points were used in the search.
A total of 2038.0 data points were assigned to clusters.
A total of 138 data points remain unassigned to clusters.

Example - acorns clustering in 2D¶

Example - `acorns` clustering in 2D¶