concatenating images in numpy array python

Question

I'm working with Python and Numpy to take several images of the same pixel dimension and create an 2D array, so each row of the array represents one image and each column will represent the pixel at a certain location.

To achieve this, I have read in the image files and tried to use numpy.concatenate. The code is

#url of picture data
X_p = data.link
#list for storing the picture data
X= []

#read in the image from the url, and skip poster with 404 error
for url in X_p:
    try:
        loadimg = urllib.request.urlopen(url)
        image_file = io.BytesIO(loadimg.read())
        img = Image.open(image_file)

        #Concatenate to linearize
        X.append(np.concatenate(np.array(img)))

    #404 error
    except urllib.error.HTTPError as err:
        if err.code == 404:
            continue
        else:
            raise

#cast the list into numpy array    
X = np.array(X)
#test to see if X is in correct dimension
print(X.shape)

I ran this code and the shape of X comes out in this format every single time

(number of images, height X width, 3)

for instance, if I load 12 image urls of 200x200 pixels, the outcome is

(12, 40000, 3)

What I need is to get rid of the 3 at the end, and it's difficult when I do not even understand where the 3 comes from.

I assume the problem I have is appending or concatenating at the wrong place. when I removed the np.concatenate, it simply did showed (12, 200, 200, 3).

I've searched online for numpy image processing and concatenations but I did not run across anything that would explain and fix what's happening.

Any and all help is appreciated. Thank you in advance for spending the time to read this post and answering..

Ishiro Kusabi · Accepted Answer · 2017-09-15 21:36:46Z

0

I figured out the problem. I was curious with dimension of my array, so I search SO for questions asking incrementing or decrementing 1 dimension. and I ran across a post that explained what the 3 stood for.

How can I save 3D array results to a 4D array in Python/numpy?

Image.open().convert("L")

did not work for me, so I had to use a trick

with Image.open().convert("L") as img

I added this line after the for loop, and the dimension problem was fixed.

answered Sep 15, 2017 at 21:36

Ishiro Kusabi

2112 gold badges6 silver badges13 bronze badges

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

Collectives™ on Stack Overflow

concatenating images in numpy array python

1 Answer 1

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related