manipulate raw binary image data with python

Question

I have an raw binary image stored as a .bin file and it has some important data on it. The problem is that the color information is slightly off so I need to alter it a little bit. Is there any way I can get in and multiply the R, G, and B values by a scalar and then save it back as a raw binary file?

I am hoping to use the python imaging library to do this as I already know the basics of the image module. I need to multiply every pixel by the same value but it will be a different value for R, G, and B. I have the following code to open the file but then I don't know what to do there to alter the RGB values.

fileName = raw_input("Enter a file name: ")

with open(fileName) as f:
    im = Image.fromstring('RGB', (3032, 2016), f.read())

Let me know if you need any more information.

UPDATE:

I wrote the following code which is converting the image in a way that I would like but it gives me an error. The code is as follows:

with open(C:\Users\name\imagedata\visiblespec.bin, 'rb') as f:
    im = Image.fromstring('RGB', (3032, 2016), f.read())

im = im.convert('RGB')
r, g, b = im.split()
r = r.point(lambda i: i * (255/171))
g = g.point(lambda i: i * (255/107))
b = b.point(lambda i: i * (255/157))
out = Image.merge('RGB', (r, g, b))


out.save(C:\Users\name\imagedata\visiblespecredone.bin)

and my error is this:

Traceback (most recent call last):
  File "C:\Users\Patrick\workspace\colorCorrect\src\rawImage.py", line 18, in <module>
    im = Image.fromstring('RGB', (3032, 2016), f.read()) 
  File "C:\Python27\lib\site-packages\PIL\Image.py", line 1797, in fromstring
    im.fromstring(data, decoder_name, args)
  File "C:\Python27\lib\site-packages\PIL\Image.py", line 594, in fromstring
    raise ValueError("not enough image data")
ValueError: not enough image data

This may be a completely wrong way to edit the RGB values, I know it works for a JPEG and t may only work for a JPEG but this is what I would like to do.

16 bit unsigned integers around 11 mbs and written in hexadecimal — clifgray
– clifgray, Commented Jun 6, 2012 at 22:41
It would appear to be that your file size is either not (or not being read as) a correct multiple of RGB values. Try .frombuffer method and check the size is what you expect? — dawg
– dawg, Commented Jun 6, 2012 at 23:29
For binary data: 3032 x 2016 = 6,112,512 pixels x 3 bytes per pixel (RGB) = 18,337,536 bytes = 17.488 MB, so 11 MB sounds too small (as in not enough image data). — martineau
– martineau, Commented Jun 7, 2012 at 2:17
Indeed is does sound like a 15 or 16 bit RGB format. Unfortunately the only raw modes that PIL v1.1.7 handles (as defined in the unpack.c file) are BGR;15 or BGR;16 which it will converted to RGB888. — martineau
– martineau, Commented Jun 7, 2012 at 2:42

cgohlke · Accepted Answer · 2012-06-07 19:07:11Z

2

import numpy as np

shape = (2016, 3032)
im = np.fromfile('visiblespec.bin', 'uint16').reshape(shape)

def tile(pattern, shape):
    return np.tile(np.array(pattern, 'bool'), (shape[0] // 2, shape[1] // 2))

r = tile([[0, 0], [0, 1]], shape)
g = tile([[0, 1], [1, 0]], shape)
b = tile([[1, 0], [0, 0]], shape)

im = im.astype('float32')
im[r] *= 255 / 171.
im[g] *= 255 / 107.
im[b] *= 255 / 157.
np.rint(im, out=im)
np.clip(im, 0, 65535, out=im)
im = im.astype('uint16')

im.tofile('visiblespecredone.bin')

edited Jun 7, 2012 at 19:07

answered Jun 7, 2012 at 18:59

cgohlke

9,4372 gold badges37 silver badges40 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

dawg · Accepted Answer · 2012-06-06 22:48:19Z

0

You might want to look at ImageMagick with the Python bindings or PIL.

If you just need to read the file and manipulate the binary data, do something like:

with open(filename, 'rb') as img:
   wrd=img.read(2)
   while wrd:
       # wrd == 2 bytes == your 16 bits...

edited Jun 6, 2012 at 22:48

answered Jun 6, 2012 at 22:41

dawg

105k24 gold badges142 silver badges217 bronze badges

6 Comments

clifgray Over a year ago

yeah I'm using PIL right now but can't quite figure out how to edit specific values with it but I'm going through the documentation again

martineau Over a year ago

Doing it this way would be very slow if done in pure Python for an image of this size.

dawg Over a year ago

@martineau: maybe, but the underlying OS will buffer this more than most give credit for.

martineau Over a year ago

@drewk: Buffering isn't going to eliminate the bit manipulations and scaling of three component values needed per pixel -- not to mention those likely necessary to write the image back out (in its original format).

dawg Over a year ago

@martineau: Agreed. It is not an ideal solution, but bit-by-bit with PIL is not going to be a speed demon either. It sounds like he has an image size error in any case, and this does not really apply.

|

Collectives™ on Stack Overflow

manipulate raw binary image data with python

2 Answers 2

Comments

6 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

6 Comments

Your Answer

Sign up or log in

Post as a guest

Related