I want to add transparency image to PDF,But the effect is not very good, the image shows a gray area

[jmu201621143028]

This my code

import fitz
from PIL import Image
import io
# https://pymupdf.readthedocs.io/en/latest/page.html#Page.insert_image

input_file = "test.pdf"
output_file = "output.pdf"
water_img = "tar.png"


# conversion Transparent
img = Image.open(water_img)
img.putalpha(128)
# conversion stream
imgbuf = io.BytesIO()
img.save(imgbuf, 'PNG')

# set location
doc = fitz.open(input_file)
page = doc[0]
rect = page.rect
br = rect.br - [50.0, 50.0]
tl = rect.tl + [50.0, 50.0]
image_rectangle = fitz.Rect(tl, br)

for i in range(len(doc)):
	page = doc[i]
	# add the image
	page.insert_image(image_rectangle, stream=imgbuf)

doc.save(output_file, deflate=True)

result image

Thanks for your code, Can you help solve this problem.

[jmu201621143028]

This is my use image

[JorjMcKie]

The main problem is, that you use the same, constant transparency value for every pixel.
You should however use complete transparency (value 0) for white pixels.
So your parameter should be a byte array of width x height single values, each of them with a suitable transparency value.
If you do this:

>>> pix=fitz.Pixmap("img-target.png")
>>> pix=fitz.Pixmap(pix, 1)  # add alpha channel
>>> for i in range(pix.height):
        for j in range(pix.width):
            pixel = list(pix.pixel(j, i))  # get pixel at this position
            if pixel == [255,255,255,255]:  # is it white?
                pixel = [0,0,0,0]  # replace by fully transparent
            else:
                pixel[-1] = 128  # else set to 50% transparency
            pix.setPixel(j, i, pixel)  # replace pixel
>>> doc=fitz.open("v110-changes.pdf")
>>> page=doc[0]
>>> page.insert_image(page.rect, pixmap=pix)
>>> doc.save("y.pdf")
>>> 

This gives me this:

[jmu201621143028]

Thank you for your answer.I understand.