mcpField
O
ocrmypdf-image

ocrmypdf-image

github.com/full-stack-skillsDocuments037 installsUpdated 4w ago
Repository

About ocrmypdf-image

ocrmypdf-image is a guide to OCRmyPDF image-processing options for improving scanned pages before OCR. It is intended for people preparing skewed, rotated, noisy, or low-resolution PDFs.

Key features

  1. 01Deskew and rotate scanned pages
  2. 02Remove borders and speckles
  3. 03Clean images with unpaper
  4. 04Oversample before OCR
  5. 05Combine processing options in recipes

Use cases

  1. 01Correct skew and orientation in scans
  2. 02Reduce noise and border artifacts
  3. 03Improve low-resolution scan inputs

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/full-stack-skills/ocrmypdf-skills --skill ocrmypdf-image

After installing, type/ocrmypdf-imagein Claude Code to confirm it loaded.

Bundled filesagents/openai.yaml

Details

From the project README

OCRmyPDF — Image Processing Guide

Overview

OCRmyPDF includes powerful image processing capabilities to improve scan quality before OCR. These tools help fix skewed pages, remove noise, clean borders, and enhance readability.

For core OCR functionality, see the ocrmypdf skill. For optimization and PDF/A options, see ocrmypdf-optimize. For batch/Docker/scripting, see ocrmypdf-batch.

Deskew

Deskew corrects pages that are slightly rotated (e.g., from feed scanner skew).

# Auto deskew (recommended)
ocrmypdf --deskew input.pdf output.pdf

# Force deskew even if rotation is minimal
ocrmypdf --deskew --force-ocr input.pdf output.pdf

Rotation

Rotate pages to correct upside-down or sideways scans:

# Auto-rotate based on text orientation
ocrmypdf --rotate-pages input.pdf output.pdf

# Force rotate all pages
ocrmypdf --rotate-pages --force-ocr input.pdf output.pdf

Remove Borders / Cleaning

Remove unwanted borders, artifacts, and noise from scanned pages:

# Remove borders (dots, solid borders)
ocrmypdf --remove-bordering input.pdf output.pdf

# Combine with cleanup
ocrmypdf --remove-bordering --clean input.pdf output.pdf

Despeckle

Remove speckles and isolated noise pixels:

# Remove speckles
ocrmypdf --despeckle input.pdf output.pdf

# Aggressive despeckle for very noisy scans
ocrmypdf --despeckle --clean input.pdf output.pdf

Unpaper

unpaper provides advanced post-processing:

# Apply unpaper with default settings
ocrmypdf --unpaper input.pdf output.pdf

# Custom unpaper board options
ocrmypdf --unpaper-args "--board A4" input.pdf output.pdf

Oversampling

Increase image resolution before OCR for better accuracy:

# Oversample to 300 DPI before OCR
ocrmypdf --oversample 300 input.pdf output.pdf

# Common for low-resolution scans
ocrmypdf --oversample 400 input.pdf output.pdf

Combined Recipes

Fix a skewed scan
ocrmypdf --deskew --remove-bordering --despeckle scanned.pdf fixed.pdf
Clean up a very noisy scan
ocrmypdf --deskew --rotate-pages --despeckle --clean --oversample 300 noisy.pdf clean.pdf
Remove all artifacts
ocrmypdf --remove-bordering --unpaper --despeckle dirty.pdf clean.pdf

Quick Reference

TaskCommand
Auto deskew--deskew
Auto rotate--rotate-pages
Remove borders--remove-bordering
Remove speckles--despeckle
Unpaper--unpaper
Oversample DPI--oversample N

Troubleshooting

  • Poor OCR after cleaning: Try --oversample 300 to increase input quality.
  • Artifacts remain: Use --unpaper for aggressive cleanup.
  • Over-cleaned image: Reduce cleaning options for preserve original quality.

Comments

Markdown supported