Python ASYNC IO Image Processing

Question

I'm currently taking images from a webcam that should be processed as soon as they are captured. The code uses the cv2.videocapture() function.

Idea: As soon as an Image is captured I want to use the time the system needs to capture a new image for computation on the previous image.

To achieve this task I use the asyncio library.

First of all I define a producer function that produces n images which should be stored in a queue:

async def produce(queue, n):
    for x in range(n):
        # produce an item
        print('producing {}/{}'.format(x, n))
        # simulate i/o operation using sleep
        await asyncio.sleep(random.random())
        item = obtainImage()
        z.append(x)
        # put the item in the queue
        await queue.put(item)

After that I implemented a consumer function which waits for an entry in the queue and computes some image features (here a simple threshold):

 async def consume(queue):
        while True:
            # wait for an item from the producer
            # Load the image stored in the queue 
            item = await queue.get()
            # process the item and store value
            ret,thresh1 = cv2.threshold(item,127,255,cv2.THRESH_BINARY)

            # simulate i/o operation using sleep
            await asyncio.sleep(random.random())

            # Notify the queue that the item has been processed
            queue.task_done() 

            #Return the Processesed image
            return thresh1

To start the functions I created a run function and added it into an event loop (The run method contains a parameter of how many images i want to process):

async def run(n):
    queue = asyncio.Queue()
    # schedule the consumer
    consumer = asyncio.ensure_future(consume(queue))
    # run the producer and wait for completion
    await produce(queue, n)
    # wait until the consumer has processed all items
    await queue.join()
    # the consumer is still awaiting for an item, cancel it
    consumer.cancel()

  loop = asyncio.get_event_loop()
  loop.run_until_complete(run(10))
  print(loop)
  loop.close()

As long as I use print statements I get an async output like this:

producing 0/10

producing 1/10

consuming 0...

producing 2/10

consuming 1...

producing 3/10

consuming 2...

producing 4/10

consuming 3...

Which is exactly what I desire in terms of order. However, I can`t get back any thresholded image. What am I missing?

Thank you in advance

user_3727050 · Answer

Minimal example with global queue for results returning (tested with Python 3.5.2)

import asyncio
import random

outQ = asyncio.Queue() 

async def produce(queue, n):
    for x in range(n):
        print('producing {}/{}'.format(x, n))
        await asyncio.sleep(random.random())
        item = x
        await queue.put(item)

async def consume(queue):
    while True:
        print('consume ')
        item = await queue.get()
        thresh1 = item**2
        print("... adding %d" % thresh1)
        await outQ.put(thresh1)
        await asyncio.sleep(random.random())
        queue.task_done()


async def run(n):
    queue = asyncio.Queue()
    consumer = asyncio.ensure_future(consume(queue))
    await produce(queue, n)
    await queue.join()
    consumer.cancel()

loop = asyncio.get_event_loop()
loop.run_until_complete(run(10))
loop.close()

while not outQ.empty():
    print(outQ.get_nowait())

Output:

...
producing 9/10
... adding 64
consume 
... adding 81
consume 
0
1
4
9
16
25
36
49
64
81

Python ASYNC IO Image Processing

1 Answers