Question

I need to perform some operations on a 2D array of values read from an image, and then create an image with a resulting 2D array. I’m using python lists to represent the 2D array.

Something very odd is happening; the values in the 2D array (list of lists) appear to become “0” at some point between the two print calls I have labeled. That is, they seem to be read correctly from the image… but then somehow get set to zero.

Code:

image = Image.open("test.png").convert("L")

data = [ [255] * image.size[1] ] * image.size[0]
pix = image.load()
for x in range(0, image.size[0]):
    for y in range(0, image.size[1]):
        data[x][y] = pix[x, y]
        #data[x][y] = 77
        print "1. data[x][y] = " + str(data[x][y]) + " .vs. " + str(pix[x, y]) # Prints correct values

for x in range(0, image.size[0]):
    for y in range(0, image.size[1]):
        print "2. data[x][y] = " + str(data[x][y]) + " .vs. " + str(pix[x, y]) # Always prints "0 .vs. [correct value]"

However, if I comment out the line

data[x][y] = pix[x, y]

And uncomment:

data[x][y] = 77

Then the two print statements show that all elements in data are 77

What is going on? I’m not an expert on python, but I can’t think of any sensible reason why list values would change like that.

I have tried the following line, in case the pixel accessor is doing something wierd:

data[x][y] = 0 + int(pix[x, y])

But still get the same result. I’ve also tried using RGB images instead of greyscale.

I should make it clear that I am definitely not doing anything with data between those two print calls. The code above is exactly what I have reduced my original program to (after discovering that all my “results” image files were black).

Answer 1

You’re problem is with this line:

data = [ [255] * image.size[1] ] * image.size[0]

This creates a list of length image.size[1] filled with the value 255. Then you create image.size[0] references to the same list and pack all of those references into another list. So, when you change a[1][1], you also change a[0][1] and a[2][1] etc. because a[0],a[1],a[2], ... are references to the same list.

Here’s a quick example:

a = [[255]*10]*10
a[1][1] = 77
print (a)

The easiest workaround is:

a = [[255]*10 for _ in range(10)]

As this creates 10 new lists instead of a list of references to the same list.