First, let's take a look at the final effect of this article ：

1 Generate qr code

To generate QR code, you can directly use the ready-made python library ：qrcode:

pip install qrcode

Use qrcode Generate qr code ：

def mk_qr(txt, dst):
img = qrcode.make(txt)
img.save(dst)

A simple test ：

mk_qr("hello world !", "text.jpg")

The generated QR code is as follows , Scan the code to see hello world ! word ：

2 gif Image analysis and generation

To generate a dynamic graph, you need to read the original gif And generate a new gif chart , That is, the function of reading and saving is required . Use opencv Can easily read gif, First installation opencv library ：

pip install opencv-python

Read gif Each frame ：

def parse_gif(path):
frames = []
cap = cv2.VideoCapture(path)
ret, image = cap.read()
while ret:
frames.append(image)
ret, image = cap.read()
cap.release()
return frames

Use imageio You can easily convert multiple frames of pictures into gif chart , First installation imageio:

pip install imageio

Save multi frame picture combination as gif：

def save_gif(frames, dst, fps=60):
imageio.mimsave(dst, frames, fps=fps)

3 Composite single frame code

Extract QR code mask

Through the QR code, we can know which position pixel values are coding areas , What are the backgrounds . We need to set the coding area to 1, The background area is set to 0.

def load_qr(path):
qr = cv2.imread(path)
qr = cv2.cvtColor(qr, cv2.COLOR_BGR2GRAY)
# Set the coding area to 1, The background area is set to 0
_, qr = cv2.threshold(qr, 240, 1, 1)
qr = cv2.cvtColor(qr, cv2.COLOR_GRAY2BGR)
return qr

We have a QR code ready , Get the mask shown below ：

Fusion frame picture

Get one ready gif chart

adopt parse_gif Function gets every frame , Fuse each frame ：

# Adjust picture brightness ,max_v Is the brightness value , The maximum is 1.0, The minimum is 0.0
def proc_frame(frame, max_v=1.0):
gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
gray = gray / 255.0
scale = max_v / (np.max(gray) + 1e-8)
frame = np.clip(frame * scale, 0 , 255).astype(np.uint8)
return frame
def mix(qr, frame):
h, w = qr.shape[0:2]
bg = np.ones((h, w, 3), np.uint8) * 255
frame = cv2.resize(frame, (w, h))
frame = proc_frame(frame)
out = qr * frame + (1 - qr) * bg
out = np.clip(out, 0, 255)
out = cv2.cvtColor(out, cv2.COLOR_BGR2RGB).astype(np.uint8)
return out

In order to prevent the QR code from being recognized because the frame is too white , You can call proc_frame Function to specify the maximum brightness value max_v, If gif Biased quantity , You can dim the overall brightness , Avoid that subsequent QR codes cannot be recognized .

It is recommended to use dark or brightly colored gif, Avoid using white or gray gif, Especially in the four corners , Avoid white .

Full code focus 【Python Learning from actual combat 】 official account , reply 2201 Get the complete code .

Welcome to follow me 【Python Learning from actual combat 】, Learn a little every day , Make a little progress every day .