How to detect non-transparent shape corner of a png image with Python?

To detect the non-transparent corners of a PNG image in Python, you can use the Python Imaging Library (PIL), now known as Pillow. Pillow allows you to load and process images, including transparent PNG images, and detect the corners based on their transparency.

Here's a step-by-step guide to detecting the non-transparent corners of a PNG image using Pillow:

Step 1: Install Pillow If you haven't already installed Pillow, you can do so using pip:

bashCopy code
pip install pillow

Step 2: Load the PNG Image Load the PNG image using Pillow's Image class:

pythonCopy code
from PIL import Image

# Load the PNG image
image = Image.open('path_to_your_image.png')

Step 3: Convert the Image to RGBA Mode If the image is not already in RGBA mode (with an alpha channel for transparency), convert it to RGBA:

pythonCopy code
if image.mode != 'RGBA':
    image = image.convert('RGBA')

Step 4: Detect the Non-Transparent Corners Now, you can analyze the alpha channel of the image to detect the non-transparent corners. Iterate through the pixels and check if the alpha value of each corner pixel is non-zero (indicating non-transparency).

pythonCopy code
def detect_non_transparent_corners(image):
    width, height = image.size
    alpha_data = image.getdata(3)  # Get alpha channel data (index 3)

    # Top-left corner
    top_left = None
    for y in range(height):
        for x in range(width):
            if alpha_data[y * width + x] > 0:
                top_left = (x, y)
                break
        if top_left:
            break

    # Top-right corner
    top_right = None
    for y in range(height):
        for x in range(width - 1, -1, -1):
            if alpha_data[y * width + x] > 0:
                top_right = (x, y)
                break
        if top_right:
            break

    # Bottom-left corner
    bottom_left = None
    for y in range(height - 1, -1, -1):
        for x in range(width):
            if alpha_data[y * width + x] > 0:
                bottom_left = (x, y)
                break
        if bottom_left:
            break

    # Bottom-right corner
    bottom_right = None
    for y in range(height - 1, -1, -1):
        for x in range(width - 1, -1, -1):
            if alpha_data[y * width + x] > 0:
                bottom_right = (x, y)
                break
        if bottom_right:
            break

    return top_left, top_right, bottom_left, bottom_right

# Call the function to get the non-transparent corners
top_left, top_right, bottom_left, bottom_right = detect_non_transparent_corners(image)

After running the code above, the variables top_left, top_right, bottom_left, and bottom_right will hold the coordinates of the non-transparent corners of the PNG image. If any of the corners are fully transparent, the corresponding variable will be None.

Note that this approach assumes the corners are defined by fully transparent pixels. If your image has partial transparency or anti-aliasing at the corners, you may need to adjust the detection logic accordingly.