In this post, I want to share how to convert PDF to images using the pdftoppm command line tool.

# Install

## Install on Windows

pdftoppm is provided by the poppler project. For Windows, in order to install the latest version of poppler, you can install it via conda:

conda install -c conda-forge poppler


On Windows, the pdftoppm tool will be installed in ANACONDA_ROOT/Library/bin. You should add this directory to the Windows PATH.

You need to install newer version of pdftoppm to use some of its features, for example, exporting to JPEG format1. Note that the poppler provided by this page is too old.

## Install on Ubuntu

apt-get update && apt-get install -y poppler-utils


This package installs the poppler command line utilities, such as pdftoppm, which we are going to use.

# How to use

To convert a single page of PDF to image, you can use the following command:

pdftoppm -singlefile -f 4 -r 72 -jpeg -jpegopt quality=90 presentation.pdf test_poppler


The PDF file we want to convert to images is presentation.pdf. The generated image name prefix is test_poppler. The image extension is decided by the exported image format. A little explanation of the options:

• -singlefile: only convert one page of PDF. It is used together with the -f option to convert a single PDF page.
• -f: index of the PDF page you want to convert. The page index starts at 1.
• -r: image DPI in both x and y direction. If you want to set DPI in x and y direction separately, use -rx and -ry instead.
• -jpeg: convert PDF page to JPEG format.
• -jpegopt: option used when convert PDF pages to JPEG images. For options and their meanings, see here.

According to my test, the pdftoppm command works great and is quick to produce the need images.

# Using pdf2image

If you want to use Python, there is also a pdf2image which is a thin wrapper around pdftoppm. Make sure you have installed pdftoppm and set its PATH correctly.

In the following script, I show an example on how to use the package.

from pdf2image import convert_from_path

def main():
pages = convert_from_path("presentation.pdf", first_page=2,
single_file=True)
pages[0].save("test_pdf2image.jpg", quality=85)

if __name__ == "__main__":
main()


The function convert_from_path() will convert the PDF to a list of PIL Image object. You can then manipulate the images with the powerful functionality provided by the Pillow package.

There also a few important parameters to note:

• dpi: this change the size and quality of th generated images. If you want to generate high quality images, use a large dpi, e.g., 300.
• thread_count: Use multi-threading to accelerate image generation. The author suggests no more than 4 threads, however, I found more threads lead to lightly faster speed. You may tweak it to fit your need.

I have also written a more detailed script to directly generate images from PPT file on the command. You can find the script here.

# References

1. Note that older version of pdftoppm only support PPM and PNG format. Newer versions support exporting to JPEG and TIFF format image. You should check whether exporting to JPEG is supported by using pdftoppm --help in the command line. ↩︎