Extract Images from a PDF

Will the extracted images lose quality?

No. The tool saves each image as it is stored in the file instead of taking a screenshot of the page. In our test, an extracted JPEG photo was byte-for-byte identical to the one placed in the PDF. A small logo stretched across a page comes out at its small stored size, so it can look blurry on its own.

Why didn't my chart or logo come out?

Charts, diagrams and logos drawn as shapes are not stored as pictures, so there is nothing to extract. If a PDF has only text and shapes, you will see No Images Found. Convert the page to PNG instead to capture them.

Saved, not screenshotted

The extractor runs pdfimages -all from the Poppler PDF library. It does not draw the page. It finds every image object the PDF stores and writes out the pixel data as it is held in the file. That is why an extracted image can look different from what you see on screen: the viewer scales, crops and layers each image when it draws the page, and none of that is applied here.

For JPEG photos the result is an exact copy. In our test, an extracted photo was byte-for-byte identical to the JPEG that had been placed in the PDF, so no quality is lost and none is added.

File types you may receive

Images keep their native encoding wherever pdfimages can write it, so one ZIP can contain several file types.

How pdfimages -all saves each kind of embedded image

Why the image size differs from the page

A PDF stores an image once at its own pixel size and then tells the page how large to draw it. The two can be far apart. In one test file, a 2480 x 1754 pixel photo was drawn only 60 mm wide, which pdfimages reported as 1050 pixels per inch on the page. It was extracted at the full 2480 x 1754 pixels. The reverse also happens: a small logo stretched across a page is extracted at its small stored size, so it can look blurry when opened alone.

Cropping is not applied either. If a document shows only part of a photo, the extracted file contains the whole stored picture, including the parts hidden on the page.

Transparency, masks and scans

• A transparent image is stored as a colour image plus a separate grayscale mask, and you receive both as separate files. In our test the transparent area of the logo came out black in the colour file, because that is the colour stored under the mask.
• Scanned PDFs usually hold one image per page. Our 3-page grayscale scan gave three 2480 x 3508 pixel single-channel JPEG files, one per page, with no text.
• Small decorative pieces, such as icons, rules drawn as images or tiny tiles, are extracted too, so expect more files than the pictures you noticed.

When nothing, or the wrong thing, comes out

• Failed to process PDF. The file may be corrupted or password-protected. pdfimages could not open the file; with our password-protected test file it stopped with an incorrect password error. Unlock the PDF first.
• The same picture appears several times: it is used on several pages. Delete the duplicates after downloading.
• A file will not open: it is probably JBIG2 or CCITT data, saved in its original form so nothing is lost. Converting the page to an image gives you a PNG of the same content.

Batch limits

You can upload up to 3 PDFs at once on the Free plan and up to 50 on Pro and AI. Each file counts as one daily check. On the Free plan, with 100 checks a day, a batch that needs more checks than you have left is refused before any work starts. Extraction of a single file stops after 2 minutes.

pdfimages, Acrobat or a screenshot instead

This tool runs Poppler's pdfimages, so if Poppler is installed you can get identical files locally: pdfimages -all report.pdf out/img. The local run has advantages: its -list option shows every image's size, color space and resolution before you extract anything, -f and -l limit it to a page range, and a sensitive document never leaves your machine. For dozens of files, a short loop in a shell is faster than any batch upload.

Acrobat Pro's Export All Images converts on the way out, so you can ask for JPEG or PNG of everything, and it can skip images below a chosen size, which removes the icons and rules this tool returns. Preview and Chrome cannot extract embedded images; copying a selection or taking a screenshot gives you the picture at screen resolution, not the stored pixels.

PDFCheck is the right pick when nothing can be installed, when you want the exact bytes of the stored JPEG with no re-encoding, or when up to 3 files on Free or 50 on Pro and AI should run as one batch from a browser.

Before you extract images

• What do I download? A ZIP named extracted_images.zip holding files called img-000, img-001 and so on, with the extension pdfimages chose for each. Numbering follows the order images appear in the file, not page numbers, and the results page also lets you open each file on its own.
• Why was my batch refused when I still had checks left? A batch is refused up front when it has more files than checks remaining, and the Free plan caps a batch at 3 files. Each file that succeeds uses one check; a file that fails uses none and is listed with its error while the others continue.
• Why did a large scan stop with the corrupted-or-protected message? pdfimages is given two minutes per file. A scanned book with hundreds of full-page images can run past that, and the tool then returns "Failed to process PDF. The file may be corrupted or password-protected." even though the file is fine. Split the PDF into parts and extract each.

How long do the images stay online?

The uploaded PDF is deleted when extraction finishes. The images and the ZIP stay until the hourly cleanup removes files older than 24 hours. Only your browser session can open the results page and the ZIP, but each image also has a plain address that anyone with the exact link can open until then.