Document scanning guide
How to Scan a Document With Your Phone and Clean It Up
A phone can capture more than enough detail for many everyday documents, but a raw camera photo rarely looks like a clean scan. The page may be trapezoidal, the desk may be visible around it, shadows may cross the paper and text contrast may vary. A good scanning workflow fixes those problems in a predictable order instead of trying to solve everything with one filter.
Start with a better source photo
Perspective correction can fix a page photographed at an angle, but it cannot restore corners that were cut out of the original photo. Keep the entire document inside the frame and leave a small border around every edge so the page boundary is easy to identify.
Hold the phone as close to parallel with the page as practical. A slight angle is fine, but extreme perspective means the far edge contains fewer useful pixels. Use even light from the sides or from a diffuse source and watch for shadows from your hand or the phone itself.
- Capture the full page, including all four corners.
- Keep the camera reasonably parallel to the document.
- Use enough light to keep text sharp and noise low.
- Avoid glare on glossy paper and strong shadows across important text.
Correct the four corners before applying aggressive cleanup
Perspective correction works by mapping the four document corners to a rectangular output. This removes the trapezoid shape created by photographing a page at an angle and also cuts away much of the surrounding table or desk.
Automatic corner detection is convenient, but it should be treated as a starting point. Zoom in and verify that every point is on the actual page corner. A small corner error can crop text, leave a strip of background or distort the final page more than necessary.
Use cleanup to improve readability, not to erase document detail
After the page is straight, contrast and shadow reduction can make it look more like a conventional scan. The correct strength depends on the document. A black-and-white printed page can tolerate stronger cleanup than a receipt with faint thermal printing, a page with colored stamps or a document containing photographs.
Check small text before export. If a filter turns thin letters into broken strokes or removes light handwriting, back off. The goal is a readable document, not an artificially pure white background at any cost.
Run OCR after geometry and readability are fixed
Optical character recognition works best when text lines are straight, characters are sharp and background noise is reduced. That makes the cleaned scan a better OCR source than the original angled phone photo.
Choose the OCR language that matches the document and review the result afterward. OCR can confuse similar characters, punctuation and numbers, especially in low-resolution or stylized text. Names, serial numbers, prices and dates deserve a manual check before the extracted text is reused.
Combine multiple cleaned pages only after each page is ready
For a multi-page document, correct and clean each page first, then combine the finished images into a PDF in the correct reading order. This is easier than creating a PDF from raw photos and trying to repair page geometry afterward.
Choose A4 or Letter when the file is meant for conventional printing, or use image-sized pages when preserving each page's natural proportions is more important. If the PDF becomes too large, reduce source image dimensions or compression before combining them.
Be careful with document privacy
Documents can contain names, addresses, signatures, identification numbers and other information that you may not want to upload to an unknown processing service. A local browser workflow lets the perspective correction, cleanup and OCR steps happen on the device instead of requiring the source image to be sent to a remote image-processing endpoint.
Also remember that image files can contain metadata from the camera or phone. If the document image will be shared directly, consider removing unnecessary metadata after the visual processing is complete.
Recommended phone-to-PDF workflow
For repeatable results, keep the order simple and do not add more processing than the document needs.
- Photograph each page with all four corners visible.
- Correct page boundaries and perspective in Document Scanner.
- Adjust cleanup only enough to improve readability.
- Run OCR on the cleaned image if you need editable text.
- Combine the final page images into a PDF and verify page order before downloading.
- Remove metadata if the original camera information is unnecessary for the shared copy.
Try the tool
Document Scanner
Apply the workflow from this guide directly in your browser.
Scan a document photo now