Hubble eXtrem Deep Field¶

Hubble eXtrem Deep Field, or HXDF, is a region located in the Fornax constellation (“Fourneau” in french), covering one thirteen-millionth of the sky, and which contains about 10,000 objects.

This shot is made from 2,000 images, is the result of 10 years of work.

Futher informations

HXDF in 3D

Setup¶

%matplotlib nbagg
from skimage import data
import numpy as np
from matplotlib import pyplot as plt
from matplotlib import cm
import numpy as np
from mpl_toolkits.mplot3d import Axes3D
from skimage.morphology import convex_hull_image
from skimage import measure

Open the image¶

image = data.hubble_deep_field()[0:500, 0:500]
channels = ['red', 'green', 'blue']

fig = plt.figure("HDF")
plt.imshow(image)
plt.axis("OFF")
plt.show()

<IPython.core.display.Javascript object>

Channel identification¶

The image has 5 informations per pixel, the x and y position and the pixel value in the 3 RGB channels : \(pixel(x,y)=[R,G,B]\)

fig = plt.figure("Channels")
for i in range(0,3):
    fig.add_subplot(1, 3, i+1)
    plt.imshow(image[:,:,i], origin = "upper", cmap = cm.gray)
    plt.title("{0}".format(channels[i]))
    plt.axis("OFF")
plt.show()

<IPython.core.display.Javascript object>

Plot the histogram¶

bins=np.arange(256)
fig = plt.figure("Histograms")
for i in range(0,3):
    fig.add_subplot(1, 3, i+1)
    plt.hist(image[:,:,i].flatten(), bins, histtype = "stepfilled",
             color="{0}".format(channels[i]))
    plt.ylim([0,18000])

    plt.title("{0}".format(channels[i]))
    plt.grid()
    plt.xlabel("Pixel value")
    plt.ylabel("Pixel count")
plt.show()

<IPython.core.display.Javascript object>

Thresholding the channels¶

fig = plt.figure("Thresholding")

for i in range(0,3):
    fig.add_subplot(1, 4, i+1)
    plt.imshow(np.where(image[:,:,i]<100,np.nan,image[:,:,i]), origin = "upper")
    plt.title("{0}".format(channels[i]))
    plt.ylim([380,420])
    plt.xlim([100,150])
    plt.axis("OFF")

fig.add_subplot(1, 4, 4)
plt.title("R+G+B")

plt.imshow(image[:,:,:], origin = "upper")
plt.ylim([380,420])
plt.xlim([100,150])
plt.axis("OFF")

plt.show()

<IPython.core.display.Javascript object>

fig = plt.figure("Star")

for i in range(0,3):
    fig.add_subplot(1, 4, i+1)
    plt.imshow(np.where(image[:,:,i]<100,0,image[:,:,i]), origin = "upper")
    plt.title("{0}".format(channels[i]))
    plt.ylim([140,190])
    plt.xlim([230,280])
    plt.axis("OFF")

fig.add_subplot(1, 4, 4)
plt.title("R+G+B")

plt.imshow(image[:,:,:], origin = "upper")
plt.ylim([140,190])
plt.xlim([230,280])
plt.axis("OFF")

plt.show()

<IPython.core.display.Javascript object>

3D plot for fun¶

image3D = data.hubble_deep_field()[380:420,100:150][:,:,1]

fig = plt.figure("3D plot of a galaxy")
ax = fig.add_subplot(111, projection='3d')

x, y = np.meshgrid(np.arange(image3D.shape[1]),
                   np.arange(image3D.shape[0]))

ax.plot_trisurf(x.flatten(), y.flatten(), image3D.flatten(), cmap=plt.cm.jet,antialiased=True)
ax.set_xlabel('x')
ax.set_ylabel('y')
ax.set_zlabel('pixel value')
plt.show()

<IPython.core.display.Javascript object>

Convex hull¶

fig = plt.figure("Convex Hull")
plt.imshow(image, interpolation='nearest', cmap=plt.cm.gray)

r= convex_hull_image(np.where(image[:,:,0]==255,1,np.nan))
g= convex_hull_image(np.where(image[:,:,1]==255,1,np.nan))
b= convex_hull_image(np.where(image[:,:,2]==255,1,np.nan))

rc=measure.find_contours(r, 0.8)
gc=measure.find_contours(b, 0.8)
bc=measure.find_contours(g, 0.8)

for n, contour in enumerate(rc):
    plt.plot(contour[:, 1], contour[:, 0], c='r', linewidth=2)
for n, contour in enumerate(gc):
    plt.plot(contour[:, 1], contour[:, 0], c='g', linewidth=2)
for n, contour in enumerate(bc):
    plt.plot(contour[:, 1], contour[:, 0], c='b', linewidth=2)
ax.axis('image')
ax.set_xticks([])
ax.set_yticks([])
plt.show()

<IPython.core.display.Javascript object>