How to convert image file object to numpy array in with openCv python?

Question

I am developing API to upload Image using Flask. After uploading i want to modify image using openCV, As openCV requires numpy array object to i have file object, how do i convert to numpy array? Here is my code

@app.route('/upload', methods=['POST'])
def enroll_user():
    file = request.files['fileName']
    # file.save(os.path.join(file.filename))
    return response

Edit: updated code

@app.route('/upload', methods=['POST'])
    def enroll_user():
        file = request.files['fileName']
        response = file.read()
        # file.save(os.path.join(file.filename))
        return response

I want to convert file to cv2 frame as i get with below code

ret, frame = cv2.imread(file)

One way is to write image to disk and read again with cv2.imread but i don't want to do that because it will be time consuming. So is there any way to convert to cv2 frame from file object?

Thanks

OpenCV image object is numpy object, generally with dtype np.uint8. — Quang Hoang
– Quang Hoang, Commented Sep 24, 2019 at 16:51
@QuangHoang You are right, OpenCV image object is numpy object, but i want to convert uploaded jpg file to OpenCV image object. — Krunal Sonparate
– Krunal Sonparate, Commented Nov 11, 2019 at 13:12

Mark Setchell · Accepted Answer · 2019-09-25 06:06:30Z

4

If you effectively have the contents of a JPEG/PNG file in your variable called response, I think you can do:

frame = cv2.imdecode(response)

Or, you may need to do:

frame = cv2.imdecode(np.fromstring(response, np.uint8), cv2.IMREAD_COLOR)

Failing that, another way is like this:

from io import BytesIO
from scipy import misc

frame = misc.imread(BytesIO(response))

edited Sep 25, 2019 at 6:06

answered Sep 24, 2019 at 21:17

Mark Setchell

210k32 gold badges309 silver badges503 bronze badges

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

5 Comments

Krunal Sonparate Over a year ago

Thanks for your reply, I have tried with imdecode, but it requires flag as second argument so i have passed cv2.IMREAD_COLOR but now i am getting Error: frame = cv2.imdecode(file, cv2.IMREAD_COLOR) TypeError: buf is not a numpy array, neither a scalar

Krunal Sonparate Over a year ago

I also tried your second method,which gives me a following Error: frame = misc.imread(BytesIO(img)) TypeError: a bytes-like object is required, not 'FileStorage'

Mark Setchell Over a year ago

I updated my answer, please try new, second option. Thank you.

Krunal Sonparate Over a year ago

Thanks, the updated option works for me, but before that i need to read the file which i haven't done. As in my updated question.

Mark Setchell Over a year ago

Cool. Glad we got you up and running - sorry it took me so long. Good luck with your project!

T.H. · Accepted Answer · 2021-07-11 16:57:31Z

0

If the file is in read-byte mode (which means mode='rb') in your case, then numpy.frombuffer() also works, for example :

file = request.files['fileName']
frame = cv2.imdecode(np.frombuffer(file.read(), numpy.uint8), cv2.IMREAD_COLOR)

answered Jul 11, 2021 at 16:57

T.H.

87610 silver badges20 bronze badges

Collectives™ on Stack Overflow

How to convert image file object to numpy array in with openCv python?

2 Answers 2

5 Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

5 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related