Error decode UTF-8 character 'â'

[tamtr1997]

I have a problem when I try using pyzbar to decode a QR image. But I had given result don't match data which I using qrcode make before.
this is my code:

from pyzbar.pyzbar import decode
from PIL import Image
from pyzbar.pyzbar import ZBarSymbol
import qrcode
image_path = "my_image.png"
data = 'â'
print(f'data = {data}')
img = qrcode.make(data)
img.save(image_path)
result = decode(Image.open(image_path),symbols=[ZBarSymbol.QRCODE])[0].data.decode("utf-8")
print(f"result = {result}")

[Mrpand4]

I'm having the same Issue using it with ã and ç, it seems to be a issue with latin characters
image_binary = take_photo()

nparr = np.frombuffer(image_binary, np.uint8)
img = cv2.imdecode(nparr, cv2.IMREAD_COLOR)

from PIL import Image

#img = cv2.imread("/content/qrcode lid.png")
barcodes = decode(img)

**** The QR code is the word Associação but pyzbar will return: data=b'Associa\xe7\x99\x9f\xe7\x93\x8ao *****
Thinking it was and encode-decode thing i tried doing several decodes but none return the right result, here is 2 of the most common encodings:

for barcode in barcodes:
try:
text = barcode[0].decode("latin-1") will return Associa��o
text = barcode[0].decode("utf-8") will return Associa癟瓊o
print(text)
except UnicodeDecodeError:
pass

[Mrpand4]

I get no issue using the qreader module so if anyone has issues with special characters I suggest using it until this module is updated

[nisanursahiner]

`from pyzbar.pyzbar import decode
from PIL import Image

image = Image.open("./test.png")
decoded_objects = decode(image)

for obj in decoded_objects:
print(obj.data.decode('utf-8').encode('shift-jis').decode('utf-8'))` You can try like this.

[luisdiegov]

I am getting a similar error but for Spanish special characters like Ñ and characters with accents like Ó or É and this solution doesn't work.

`from pyzbar.pyzbar import decode from PIL import Image

image = Image.open("./test.png") decoded_objects = decode(image)

for obj in decoded_objects: print(obj.data.decode('utf-8').encode('shift-jis').decode('utf-8'))` You can try like this.

[YDX-2147483647]

This issue has been repeatedly reported: #55, #95, #155, …

The cause is that ZBar guesses the encoding of the data, but for short texts, the guess is usually wrong.

Experiments

from pyzbar.pyzbar import decode
from qrcode import make

def round_trip(a: bytes) -> bytes:
    return decode(make(a).get_image())[0].data


def f(src: list[int], dst: list[int]) -> bool:
    return round_trip(bytes(src)) == bytes(dst)


assert all(
    [
        *(f([i], [i]) for i in range(0x80)),
        *(f([i], [0xC2, i]) for i in range(0x80, 0xA1)),
        *(f([i], [0xEF, 0xBD, i]) for i in range(0xA1, 0xC0)),
        *(f([i], [0xEF, 0xBE, i - 0x40]) for i in range(0xC0, 0xE0)),
        *(f([i], [0xC3, i - 0x40]) for i in range(0xE0, 0x100)),
        round_trip(b"\xe5\x95") == b"\xe8\x9d\xa0",
        round_trip(b"\xe5\x95\x8a") == b"\xe5\x95\x8a",
        round_trip(b"\xe2\x80\xa6") == b"\xe7\xaa\xb6\xef\xbd\xa6",
        round_trip(b"\xe5\x95\x8a\xe2\x80\xa6") == b"\xe5\x95\x8a\xe2\x80\xa6",
    ]
)

Workaround if you only handle texts

Use OpenCV's QRCodeDetector.

import cv2
import numpy as np
from qrcode import make

detector = cv2.QRCodeDetector()

def cv_round_trip(a: str) -> str:
    arr = np.array(make(a.encode("utf-8")).get_image(), dtype=np.uint8) * 0xff
    return detector.detectAndDecode(arr)[0]

Workaround if you also handle binary data

Maybe #82 can fix it.