Reading a barcode using OpenCV QRCodeDetector - python-3.x

Question

I am trying to use OpenCV on Python3 to create an image with a QR code and read that code back.
Here is some relevant code:
def make_qr_code(self, data):
qr = qrcode.QRCode(
version=2,
error_correction=qrcode.constants.ERROR_CORRECT_H,
box_size=10,
border=4,
)
qr.add_data(data)
return numpy.array( qr.make_image().get_image())
# // DEBUG
img = numpy.ones([380, 380, 3]) * 255
index = self.make_qr_code('Hello StackOverflow!')
img[:index.shape[0], :index.shape[1]][index] = [0, 0, 255]
frame = img
# // DEBUG
self.show_image_in_canvas(0, frame)
frame_mono = cv.cvtColor(numpy.uint8(frame), cv.COLOR_BGR2GRAY)
self.show_image_in_canvas(1, frame_mono)
qr_detector = cv.QRCodeDetector()
data, bbox, rectifiedImage = qr_detector.detectAndDecode(frame_mono)
if len(data) > 0:
print("Decoded Data : {}".format(data))
self.show_image_in_canvas(2, rectifiedImage)
else:
print("QR Code not detected")
(the calls to show_image_in_canvas are just for showing the images in my GUI so I can see what is going on).
When inspecting the frame and frame_mono visually, it looks OK to me
However, the QR Code Detector doesn't return anything (going into the else: "QR Code not detected").
There is literally nothing else in the frame than the QR code I just generated. What do I need to configure about cv.QRCodeDetector or what additional preprocessing do I need to do on my frame to make it find the QR code?

Answer 1

OP here; solved the problem by having a good look at the generated QR code and comparing it to some other sources.
The problem was not in the detection, but in the generation of the QR codes.
Apparently the array that qrcode.QRCode returns has False (or maybe it was 0 and I assumed it was a boolean) in the grid squares that are part of the code, and True (or non-zero) in the squares that are not.
So when I did img[:index.shape[0], :index.shape[1]][index] = [0, 0, 255] I was actually creating a negative image of the QR code.
When I inverted the index array the QR code changed from the image on the left to the image on the right and the detection succeeded.
In addition I decided to switch to the ZBar library because it's much better at detecting these codes under less perfect circumstances (like from a webcam image).

Answer 2

import cv2
import sys
filename = sys.argv[1]
# Or you can take file directly like this:
# filename = f'images/filename.jpg' where images is folder for files that you trying to read
# read the QRCODE image
# in case if QR code is not black/white it is better to convert it into grayscale
# Zero means grayscale
img = cv2.imread(filename, 0)
img_origin = cv2.imread(filename)
# initialize the cv2 QRCode detector
detector = cv2.QRCodeDetector()
# detect and decode
data, bbox, straight_qrcode = detector.detectAndDecode(img)
# if there is a QR code
if bbox is not None:
print(f"QRCode data:\n{data}")
# display the image with lines
# length of bounding box
# Cause bbox = [[[float, float]]], we need to convert fload into int and loop over the first element of array
n_lines = len(bbox[
0])
bbox1 = bbox.astype(int) # Float to Int conversion
for i in range(n_lines):
# draw all lines
point1 = tuple(bbox1[0, [i][0]])
point2 = tuple(bbox1[0, [(i + 1) % n_lines][0]])
cv2.line(img_origin, point1, point2, color=(255, 0, 0), thickness=2)
# display the result
cv2.imshow("img", img)
cv2.waitKey(0)
cv2.destroyAllWindows()
else:
print("QR code not detected")

Answer 3

To re-state the accepted answer, the background of the QRcode must be white and the foreground must be black. So if the generated QRcode has a white foreground you must invert the colors, e.g.:
from cv2 import cv2
img = cv2.imread('C:/Users/N/qrcode.jpg')
img_gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
# Invert colors so foreground is black
img_invert = cv2.bitwise_not(img_gray)
cv2.imshow('gray', img_gray)
cv2.imshow('inverted', img_invert)
cv2.waitKey(1)
qr_detector = cv2.QRCodeDetector()
text, _, _ = qr_detector.detectAndDecode(img_invert)
print(text)