python - How detect and plot intensity of asc file -

i have photo of flame - data asc file contains matrix of pixels. in each pixel value of light intensity.

my code of plotting:

import os import numpy np import matplotlib.pyplot plt # open file path = "input/" dirs = os.listdir( path ) number_of_files = 0  # print files , directories file in dirs:     if file.endswith(".asc"):         img = np.genfromtxt (path+file)          file = os.path.splitext(file)[0]         #pasukamas vaizdas         img = img.transpose ()         # figure , 1 subplot         fig, ax = plt.subplots(1,figsize=(20,20))          ax.set_title(file, fontsize=60, y=1.03)         plt.imshow (img, interpolation='nearest', origin='lower')          plt.colorbar ()         plt.savefig ('output/' + file + '.png', bbox_inches = 'tight')          number_of_files = number_of_files + 1         plt.close(fig) print (number_of_files) 

result:

how can plot images in 3 ranges:

1. from max intensity 36000 (red zone)
2. from max intensity 27000 (yellow zone)
3. from max intensity 12000 (blue zone)

and detect bottom , top pixels? left , right pixels? connect top , bottom pixels line. need same thing left , right pixels. how show pixel distance on lines?

result must

you seem want thresholded version of image(s), area labeling on thresholded values, , fancy measures afterwards.

---

for convenience form 3d ndarray of image sequences , operation can done in 1 go numpy :

filelist = filter(lambda s: s.endswith(".asc"), os.listdir(path))  # open first image have dimensions initialize array firstimage = np.genfromtxt (path+filelist[0])  imagearray = np.zeros((len(filelist),) + firstimage.shape) 

now assign values

imagearray[0,:,:] = firstimage i,file in enumerate(filelist[1:]):     # skip first item because have     imagearray[i+1,:,:] = np.genfromtxt (path+file) 

---

ok, have 3d array of image, let's range images

boolmaskredzone = imagearray > 36000  boolmaskyellowzone = imagearray > 27000  boolmaskyellowzone = imagearray > 12000  

these masks of same size images of booleans. let's fiddle :

redparts = image*boolmaskredzone # images 0 thresholded values plt.imshow(redparts[0,:,:],cmap="hot") 

notice again redparts , else still in 3d, made 2d view of array plotting purposes.

---

now easy / fun part : labeling ! can use scipy.ndimage.measurements.label()

from scipy.ndimage import label labelsred, nblabelsred = label(boolmaskredzone) 

labelsred array ints label indices.

ideally have nblabelsred == 1, if not, "islands" can closed with

from scipy.ndimage import morphology closedlabels = morphology.binary_closing(labelsred) # fiddle optional iterations parameter if needed 

we can compute area of labels = thresholded areas using np.where give positions of pixels, counting number of items :

x,y,z = np.where(labelsred == 1) # x, y ,z arrays area = len(x) # number of pixels brighter red 

as computing top/bottommost pixels , can become tricky if want line diagonal one, if want top / bottom (aligned images axes) can make numpy check when masks become true each axis, plainly doing difference (derivation) between array , and offset version, first nonzero element along each axis

differencearray = boolmaskredzone - np.roll(boolmaskredzone,1,axis=1) # check along each column when array first becomes true uselessimageindex,xtopmostpixel,ytopmostpixel= numpy.where(differencearray == true) # found crossings , put them in 2 arrays 

---

for exact diagonal measurement, might want specialized image measurement libraries, scikit-image, have want

if want yourself, recommend approach based on finding object centre, computing diagonal lines position, , measuring max length, happens if find 45 degree line ? become "top-bottom" or "left-right" ? or want longest line close horizontal ? or want bi-orthogonal measurement ? (pick biggest line first measurement, second diameter line whatever length of line 90 degrees first one)

assuming have points in image, plotting line plot plt.plot =)

i have admit didn't think through derivation part much, assumed once you'd have label happy camper =)

edit : of course operations can computed directly iterating on array, posted approach produces one-liners use numpy's array manipulation efficiency stuff. indeed each operation doing

for x in range(image.shape[0]):     y in range(image.shape[1]):         if(image[x,y] > 36000):             mask[x,y]=true 

but nesting of loops very slow compared numpy's compiled , hardware-accelerated functions (see http://www.astro.washington.edu/users/vanderplas/astr599/notebooks/11_efficientnumpy demo of speed on python/numpy)

edit 2: project of mine i've been researching more of scipy's ndimage functions, , there's : ndimage.center_of_mass().

this function finds center of mass in (potentially labelled) array. finding center of mass of labelled array, have center of axis find diagonals from, , rest piece of cake ^^