问题
I want to crop an image to its smaller size, by cutting the white areas on the borders. I tried the solution suggested in this forum Crop a PNG image to its minimum size but the getbbox() method of pil is returning a bounding box of the same size of the image, i.e., it seems that it doesn't recognize the blank areas around. I tried the following:
>>>import Image
>>>im=Image.open("myfile.png")
>>>print im.format, im.size, im.mode
>>>print im.getbbox()
PNG (2400,1800) RGBA
(0,0,2400,1800)
I checked that my image has truly white croppable borders by cropping the image with the GIMP auto-crop. I also tried with ps and eps versions of the figure, without luck.
Any help would be highly appreciated.
回答1:
Trouble is getbbox()
crops off the black borders, from the docs: Calculates the bounding box of the non-zero regions in the image
.
import Image
im=Image.open("flowers_white_border.jpg")
print im.format, im.size, im.mode
print im.getbbox()
# white border output:
JPEG (300, 225) RGB
(0, 0, 300, 225)
im=Image.open("flowers_black_border.jpg")
print im.format, im.size, im.mode
print im.getbbox()
# black border output:
JPEG (300, 225) RGB
(16, 16, 288, 216) # cropped as desired
We can do an easy fix for white borders, by first inverting the image using ImageOps.invert
, and then use getbbox()
:
import ImageOps
im=Image.open("flowers_white_border.jpg")
invert_im = ImageOps.invert(im)
print invert_im.getbbox()
# output:
(16, 16, 288, 216)
来源:https://stackoverflow.com/questions/9870876/getbbox-method-from-python-image-library-pil-not-working