如何将Gdk pixbuf对象转换为numpy数组 [英] How to turn Gdk pixbuf object into numpy array

查看:100
本文介绍了如何将Gdk pixbuf对象转换为numpy数组的处理方法,对大家解决问题具有一定的参考价值,需要的朋友们下面随着小编来一起学习吧!

问题描述

因此,我尝试拍摄屏幕截图,并以numpy数组的形式对其进行处理.但是,我所能找到的就是先将其转换为PIL Image,然后转换为numpy数组.因此,我想知道是否有一种方法可以在不使用PIL库的情况下直接将其转换为numpy数组

So, I am trying to take a screenshot of the screen and do some stuff with it as numpy array. However, all I could find is turning it into PIL Image first and then into numpy array. So, I was wondering if there is a way of turning it straight into numpy array without using the PIL library

from gi.repository import Gdk

window = Gdk.get_default_root_window()
x, y, width, height = window.get_geometry()
pb = Gdk.pixbuf_get_from_window(window, x, y, width, height)

谢谢

推荐答案

如果您不介意使用Gtk2和PyGtk,则可以执行以下操作

If you don't mind using Gtk2 and PyGtk, then the following will do

import gtk, numpy
image="yourimage.jpg"
t=gtk.gdk.pixbuf_new_from_file(image)
n=t.get_pixels_array()
print n.shape

最好的部分是n是对t中数据的引用,因此您可以使用numpy更改pixbuf

The best part is that n is a reference to the data in t so you can alter the pixbuf using numpy

如果您想使用Gtk3,那么这两个调用就可以了

If you wish to use Gtk3, these two calls will do

from gi.repository import GdkPixbuf
import numpy
def array_from_pixbuf(p):
    " convert from GdkPixbuf to numpy array"
    w,h,c,r=(p.get_width(), p.get_height(), p.get_n_channels(), p.get_rowstride())
    assert p.get_colorspace() == GdkPixbuf.Colorspace.RGB
    assert p.get_bits_per_sample() == 8
    if  p.get_has_alpha():
        assert c == 4
    else:
        assert c == 3
    assert r >= w * c
    a=numpy.frombuffer(p.get_pixels(),dtype=numpy.uint8)
    if a.shape[0] == w*c*h:
        return a.reshape( (h, w, c) )
    else:
        b=numpy.zeros((h,w*c),'uint8')
        for j in range(h):
            b[j,:]=a[r*j:r*j+w*c]
        return b.reshape( (h, w, c) )

def pixbuf_from_array(z):
    " convert from numpy array to GdkPixbuf "
    z=z.astype('uint8')
    h,w,c=z.shape
    assert c == 3 or c == 4
    if hasattr(GdkPixbuf.Pixbuf,'new_from_bytes'):
        Z = GLib.Bytes.new(z.tobytes())
        return GdkPixbuf.Pixbuf.new_from_bytes(Z, GdkPixbuf.Colorspace.RGB, c==4, 8, w, h, w*c)
    return GdkPixbuf.Pixbuf.new_from_data(z.tobytes(),  GdkPixbuf.Colorspace.RGB, c==4, 8, w, h, w*c, None, None)

这篇关于如何将Gdk pixbuf对象转换为numpy数组的文章就介绍到这了,希望我们推荐的答案对大家有所帮助,也希望大家多多支持IT屋!

查看全文
登录 关闭
扫码关注1秒登录
发送“验证码”获取 | 15天全站免登陆