Send and receive images using Flask, Numpy and OpenCV

client.py

from __future__ import print_function
import requests
import json
import cv2

addr = 'http://localhost:5000'
test_url = addr + '/api/test'

# prepare headers for http request
content_type = 'image/jpeg'
headers = {'content-type': content_type}

img = cv2.imread('lena.jpg')
# encode image as jpeg
_, img_encoded = cv2.imencode('.jpg', img)
# send http request with image and receive response
response = requests.post(test_url, data=img_encoded.tostring(), headers=headers)
# decode response
print(json.loads(response.text))

# expected output: {u'message': u'image received. size=124x124'}

server.py

from flask import Flask, request, Response
import jsonpickle
import numpy as np
import cv2

# Initialize the Flask application
app = Flask(__name__)


# route http posts to this method
@app.route('/api/test', methods=['POST'])
def test():
    r = request
    # convert string of image data to uint8
    nparr = np.fromstring(r.data, np.uint8)
    # decode image
    img = cv2.imdecode(nparr, cv2.IMREAD_COLOR)

    # do some fancy processing here....

    # build a response dict to send back to client
    response = {'message': 'image received. size={}x{}'.format(img.shape[1], img.shape[0])
                }
    # encode response using jsonpickle
    response_pickled = jsonpickle.encode(response)

    return Response(response=response_pickled, status=200, mimetype="application/json")


# start flask app
app.run(host="0.0.0.0", port=5000)

You don't really need OpenCV to post an image. requests alone is enough.

def post_image(img_file):
    """ post image and return the response """
    img = open(img_file, 'rb').read()
    response = requests.post(URL, data=img, headers=headers)
    return response

Sometimes this can be useful when you process an image before sending

Sometimes you might not even have a code running at the other end to imdecode which is exactly what happened with me. Where I had to upload an image directly onto google-cloud-storage
Here's a solution I came up with

from google.cloud import storage
import cv2
from tempfile import NamedTemporaryFile
import six

client = storage.Client()
#specify google bucket name
bucket = client.get_bucket('doc')

#image to be loaded
image = cv2.imread('24.jpg')

with NamedTemporaryFile() as temp:
    #add JPEG format to the NamedTemporaryFile  
    iName = "".join([str(temp.name),".jpg"])
    
    #save the numpy array image onto the NamedTemporaryFile
    cv2.imwrite(iName,image)
    #Name of the file in the bucket
    blob = bucket.blob('T15.jpg')
    #Type of file being uploaded
    blob.upload_from_filename(iName,content_type='image/jpeg')
    #get the public url
    url = blob.public_url
    print(type(url))

Thanks for this! I hit a problem where I needed to encode the image before sending and decode it again. If you convert the image into gray scale and use the received image in dlib (face_recognition) then library complains with RuntimeError: Unsupported image type, must be 8bit gray or RGB image.. Encoding and decoding to cv2.IMREAD_COLOR helped me solve this problem. Decoding the payload to IMREAD_GRAYSCALE didn't help though.

Thank you @kyle Hounslow @kylehounslow Thank you very very much !!!!

It really helped me .. Thank you again
I had similar problem @kaskavalci

What if i want to send some additional data along with the frame eg: name, id etc.

Thanks, useful advice.

It will be really nice If script can be completely standalone/independent RESTful Web Services that can process the image along with some text data to go with it. regardless to what the client side language is. So, the client side and server side are completely separated.

@ALL .........How can we send encoded image from frontend to the server ....Please anyone

Thanks for your code sample.

by the way,

I am using aiohttp lib for handling async request from multiple clients.

But I am a beginner and I don't know how to implement the same (above code sample) using in aiohttp lib (bcz to handle the multiple req at a time)

highly appreciated your help

how can i get image in flask api responce?

and how can I receive two images?

I try to send a image to Azure virtual machine and used the same code. It works in my localhost But I am getting this error: ConnectionError: ('Connection aborted.', ConnectionResetError(10054, 'An existing connection was forcibly closed by the remote host', None, 10054, None))
Do you know how can i solve this problem? @kylehounslow

Hi, I want to know how to pass some parameter in the test() function.

I want to apply sone deeplearning method in image processing, while, deeplearning method is seperate in 2 parts, because the initial part is so time consuming, and I want to pass the model created by the initial function of deeplearning into the test() function, how should I do?

Thanks a lot!

Hi, is it possible to send some more information like roi, image_name, etc. with the encoded image using the same server client method.

how can i get image in flask api responce?

frame = analyze_and_modify_frame( frame )  
_, frame  = cv.imencode('.jpg', frame)
response_pickled  = jsonpickle.encode( frame )
return Response(response=response_pickled, status=200, mimetype="application/json") 

and on receiving client:

frame = jsonpickle.decode( response.text ) 
frame = frame.tobytes()

How can we send and receive multiple images, is it possible to encapsulate images and other data in a dictionary object and send

it tooks more time to send image to flask, if i tried on localhost it works well but got issue when trying to post image to hosted app. any suggestions?

r = request
nparr = np.fromstring(r.data, np.uint8)
# decode image
img = cv2.imdecode(nparr, cv2.IMREAD_COLOR)

img = cv2.rotate(img, cv2.ROTATE_90_COUNTERCLOCKWISE)

img = cv2.imencode('.png', img)

response_pickled = jsonpickle.encode(img)

return Response(response=response_pickled, status=200, mimetype="application

How to get this image rotated back as a response? Any suggestions?
cv2.imencode function not work properly to receive image for example in postman

how can I use cv2.imshow() in server.
Please let us.

@amarnathreddy0201 you can't use cv2.imshow in server
What imshow does, is open a window and display the image in there right?
And from a server side you don't usually have a user interface ... what I would do
1: generate the image using cv2 saving it on a file o just keeping it on memory
2: Find out how to share that image to someone outside the server (Make a RESt API maybe or generate something with sockets)

Maybe this will help https://stackoverflow.com/questions/55822414/how-to-display-an-encoded-opencv-image-on-a-flask-web-server

Thanks for this @kylehounslow passing cv2.IMREAD_COLOR to cv2.imdecode was where I was missing it.[ np.fromstring ](https://numpy.org/devdocs/reference/generated/numpy.fromstring.html) is deprecated tho. np.drombuffer should be used

Anyone that wants to get image from client and store in buffer can use this snippet below for flask

       ```
        original_image = request.files.get('original image')
        compare_image = request.files.get('compare image')
        original_img = cv2.imdecode(np.frombuffer(original_image.read(), np.uint8), cv2.IMREAD_COLOR)
        compare_img = cv2.imdecode(np.frombuffer(compare_image.read(), np.uint8), cv2.IMREAD_COLOR)
        known_encoding = face_recognition.face_encodings(original_img)[0]
        unknown_encoding = face_recognition.face_encodings(compare_img)[0]

        results = face_recognition.compare_faces([known_encoding], unknown_encoding)

@kylehounslow Thanks for the this understandable scripts.
I would like to pass some more additional information along with image to server. Something like below

meta = {'camera': 'camera', 'distance': 'distance'}
response = requests.post(test_url, data=img_encoded.tostring(), headers=headers, json=meta)

And I access this meta information on the server as input_json = r.get_json(force=True)

But this results in error in client side as below.

Traceback (most recent call last):
  File "/home/keerti/PycharmProjects/Flask_Api/client.py", line 21, in <module>
    print(json.loads(response.text))
  File "/usr/lib/python3.10/json/__init__.py", line 346, in loads
    return _default_decoder.decode(s)
  File "/usr/lib/python3.10/json/decoder.py", line 337, in decode
    obj, end = self.raw_decode(s, idx=_w(s, 0).end())
  File "/usr/lib/python3.10/json/decoder.py", line 355, in raw_decode
    raise JSONDecodeError("Expecting value", s, err.value) from None
json.decoder.JSONDecodeError: Expecting value: line 1 column 1 (char 0)

Could you please help me with this? If this is the right way to pass additional information