蟒蛇&OpenCV [英] Python & OpenCV

问题描述

是否有任何本机支持从 PDF 抓取图像或在 Python 中创建某种对象，这些对象可以包含来自 pdf 的图像，然后可以通过 OpenCV 访问?我已经查看了一些脚本来将 PDF 的图像转储到我的目录中，但我的目标更多是访问 PDF，而是将 PDF 中的图像数据转储到我可以访问的某种对象中使用 OpenCV.我自己的探索没有产生任何结果，所以我想我会问.

基于@Ghilas BELHADJ 的示例添加了使用 PyMuPDF 的示例

导入菲茨导入 cv2将 numpy 导入为 np从 tkinter 导入 Tk从 tkinter.filedialog 导入 askopenfilename类 AccessPDF:def __init__(self):self.filepath = ""self.doc = 无def openPDF(self):Tk().withdraw()self.filepath = askopenfilename()self.doc = fitz.open(self.filepath)def pixel2np(self,pix):im = np.frombuffer(pix.samples, dtype=np.uint8).reshape(pix.h, pix.w, pix.n)im = np.ascontiguousarray(im[..., [2, 1, 0]]) # rgb 到 bgr返回我def displayKey(self):pixobj = self.doc.getPagePixmap(0, alpha=False)im = self.pixel2np(pixobj)cv2.imwrite("testimg.png",im)cv2.imshow("Key" im)

解决方案

我根据@Dan Mašek 的评论对代码进行了修改

您可以使用

Is there any Native Supports for Grabbing Images from PDFs or Create some sort of Object in Python that can contain the Images from a pdf that then can be access via OpenCV? I've looked at some scripts to dump the Images of a PDF into my directory but I'm aiming more at accessing the PDF and instead dumping the data from the PDF that is the image(s) into some sort of object I can access with OpenCV. My own exploration hasn't yielded any results so i figured I'd ask.

Added a Example of Using PyMuPDF based off example from @Ghilas BELHADJ

import fitz
import cv2
import numpy as np
from tkinter import Tk
from tkinter.filedialog import askopenfilename


class AccessPDF:

    def __init__(self):
        self.filepath = ""
        self.doc = None

    def openPDF(self):
        Tk().withdraw()
        self.filepath = askopenfilename()
        self.doc = fitz.open(self.filepath)

    def pixel2np(self,pix):
        im = np.frombuffer(pix.samples, dtype=np.uint8).reshape(pix.h, pix.w, pix.n)
        im = np.ascontiguousarray(im[..., [2, 1, 0]])  # rgb to bgr
        return im

    def displayKey(self):  
        pixobj = self.doc.getPagePixmap(0, alpha=False)
        im = self.pixel2np(pixobj)
        cv2.imwrite("testimg.png",im)
        cv2.imshow("Key" im)

解决方案

Edit: I've made a modification in the code following the comment of @Dan Mašek

You can achieve this (load the PDF embedded images into OpenCV without writing intermediate objects on disk) using PyMuPDF and Numpy.

In this example, I'm using this pdf file.

import fitz
import cv2
import numpy as np


def pix2np(pix):
    im = np.frombuffer(pix.samples, dtype=np.uint8).reshape(pix.h, pix.w, pix.n)
    im = np.ascontiguousarray(im[..., [2, 1, 0]])  # rgb to bgr
    return im


doc = fitz.open('NGM_2018_Media_Kit.pdf')

# entire page
# pix = doc.getPagePixmap(0, alpha=False)

# first page , 5th image, xref element
pix = fitz.Pixmap(doc, doc.getPageImageList(0)[4][0])  
im = pix2np(pix)

cv2.putText(im, 'Azul fellawen', (100, 100),
            cv2.FONT_HERSHEY_SIMPLEX, 1.,
            (18, 156, 243), 2, cv2.LINE_AA)
cv2.imwrite('sample_0.png', im)

这篇关于蟒蛇&OpenCV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！