Using Webcam with Yolov5 Models-CodePudding

I tried to run the example infer-simple.py, but I can't succeed, because the code is returning a None value when the command line (image = cv2.imdecode(image, cv2.IMREAD_COLOR)) is executed.

code:

def infer():
    # Get the current image from the webcam
    ret, img = video.read()

    # Resize (while maintaining the aspect ratio) to improve speed and save bandwidth
    height, width, channels = img.shape
    scale = ROBOFLOW_SIZE / max(height, width)
    img = cv2.resize(img, (round(scale * width), round(scale * height)))

    # Encode image to base64 string
    retval, buffer = cv2.imencode('.jpg', img)
    img_str = base64.b64encode(buffer)

    # Get prediction from Roboflow Infer API
    resp = requests.post(upload_url, data=img_str, headers={
        "Content-Type": "application/x-www-form-urlencoded"
    }, stream=True).raw

    # Parse result image
    image = np.asarray(bytearray(resp.read()), dtype="uint8")
    image = cv2.imdecode(image, cv2.IMREAD_COLOR)
    
    return image

Always before the imdecode image is array not None, bellow is displayed one debug example.

Debug

resp: <urllib3.response.HTTPResponse object at 0x0000015055DE7070>

image: array([ 11, 214,  13, ..., 170,   1,   3], dtype=uint8)

However, when I run cv2.imdecode(image, cv2.IMREAD_COLOR)) I get a none value for image.

3.Error

Exception has occurred: error OpenCV(4.6.0) D:\a\opencv-python\opencv-python\opencv\modules\imgcodecs\src\loadsave.cpp:816: 
error: (-215:Assertion failed) !buf.empty() in function 'cv::imdecode_'

  File "C:\Users\diego\codes\Webcam\infer-simple.py", line 48, in infer
    image = cv2.imdecode(image, cv2.IMREAD_COLOR)
  File "C:\Users\diego\codes\Webcam\infer-simple.py", line 63, in <module>
    image = infer()

I did it all step by step according to https://blog.roboflow.com/python-webcam. Except for putting the local host to compose the url, because when I put my local host I get an error.

Example error:

NewConnectionError('<urllib3.connection.HTTPConnection object at 0x00000277283F1E80>: 

Failed to establish a new connection: [WinError 10061]

I applied solutions found on the Internet and also reinstalled all necessary packages but nothing worked.

CodePudding user response：

Problem is not when you get image from camera but when you get image from server.

You have to create project on roboflow.com and gets model's name and api key - and use all this to create correct url.

I created project and trained model with name chess-sample-cpuhx/1 and got API KEY like tE7xxxxxxxxx so I have full url

https://detect.roboflow.com/chess-sample-cpuhx/1?format=image&stroke=5&api_key=tE7xxxxxxxxx

My code with few changes.

I check

if webcam gives image
if server send response status 200 and if status is different then display message from server - like b'{"message":"Not Found"}' or b'{"message":"Forbidden"}'

# all `import` at the beginning
import json
import base64
import cv2
import numpy as np
import requests
import time

# --- constants ---  

ROBOFLOW_API_KEY = 'tE7xxxxxxxxx'
ROBOFLOW_MODEL = 'chess-sample-cpuhx/1'
ROBOFLOW_SIZE = 400

# --- functions ---

params = {
    "api_key": ROBOFLOW_API_KEY,
    "format": "image",
    "stroke": "5"
}

headers = {
    "Content-Type": "application/x-www-form-urlencoded"
}

url = f"https://detect.roboflow.com/{ROBOFLOW_MODEL}"


def infer(img):
    # Resize (while maintaining the aspect ratio) to improve speed and save bandwidth
    height, width, channels = img.shape
    scale = ROBOFLOW_SIZE / max(height, width)
    img = cv2.resize(img, (round(scale * width), round(scale * height)))

    # Encode image to base64 string
    retval, buffer = cv2.imencode('.jpg', img)
    img_str = base64.b64encode(buffer)

    # Get prediction from Roboflow Infer API
    response = requests.post(url, params=params, data=img_str, headers=headers, stream=True)
    data = response.raw.read()
    
    #print(response.request.url)
    
    if not response.ok:
        print('status:', response.status_code)
        print('data:', data)
        return
    
    # Parse result image
    image = np.asarray(bytearray(data), dtype="uint8")
    image = cv2.imdecode(image, cv2.IMREAD_COLOR)

    return image

# --- main ---

video = cv2.VideoCapture(0)

while True:

    start = time.time()

    ret, img = video.read()
    
    if ret:
        
        image = infer(img)
        
        if image is not None:
            cv2.imshow('image', image)
        
            if cv2.waitKey(1) == ord('q'):  # `waitKey` should be after `imshow()` - to update image in window and to get key from window
                break
        
            end = time.time()
            print( 1/(end-start), "fps")  # print() automatically add space between elements - you don't need space in "fps"
        
# - end -

video.release()
cv2.destroyAllWindows()