使用 Homebrew, apk, apt, dnf, MacPorts, Nix, pacman, zypper 安装 unpaper

项目链接

• 主页
• 仓库
• 上游文档
• 软件包管理器页面

命令和别名

• unpaper

项目历史

The upstream README says unpaper was originally written by Jens Gulden, and the AUTHORS file lists Gulden as original author and maintainer from 2005-2009. The NEWS file records the first release, 1.0, on 2005-03-01.

Early releases focused on the core scanned-book workflow: joining or splitting pages, handling double-page layouts, deskewing, stretching, zooming, border alignment, black filtering, and multi-file batch processing. The 2005-2007 NEWS entries also show portability work for Mac OS X and SPARC Solaris, color support through PPM, documentation diagrams, binary/source distribution archives, and conventional tarball packaging.

Maintenance later moved from a personal scanned-page tool into a modern hosted open source project. NEWS records an Autotools build system in 2011, source reorganization and parser cleanup in 2012, mathematically accurate interpolation in 2013, libav-based file input/output in 2014, ffmpeg compatibility updates, and a 2022 Meson migration in release 7.0.0.

采用历史

The supplied package metadata shows unpaper available across a broad set of Unix-like package managers: Alpine/apk, Homebrew, Debian, Fedora/dnf, MacPorts, Nix, Arch/pacman, Ubuntu, and openSUSE/zypper. That matches its niche as a small C utility used inside scanning, OCR, and PDF-preparation pipelines.

The project's dependency history also tracks adoption pressure. The README now says ffmpeg is the only hard dependency for file input and output; NEWS shows earlier transitions through Netpbm-style PNM workflows, libav, ffmpeg compatibility, and build-system modernization.

使用方式

unpaper is usually run in batch mode over numbered image sequences. The basic concepts documentation explains sheets, pages, single-page and double-page layouts, wildcard filename sequences such as `input%03d.pbm`, and options for joining single-page scans into double-page sheets or splitting double-page scans back into individual page images.

The README describes the processing features that define its everyday use: removing dark borders from scans and photocopies, detecting misaligned centering and rotation, deskewing pages, and allowing each processing step to be disabled or tuned when automatic processing fails.

为什么软件包爱好者会关心

unpaper is significant in package-manager culture because it is one of those small, sharp Unix tools that exists to make a larger workflow possible. It sits between scanner output, image conversion tools, OCR engines, and PDF builders, so distributions care about stable command-line behavior, supported image formats, and dependency choices more than about a GUI.

It is also a good example of a long-lived niche media utility being kept buildable as the C ecosystem changes: custom code gave way to libav/ffmpeg for formats, Autotools gave way to Meson, tests were adapted for floating-point differences, and maintainers kept portability and package builds alive.

时间线

• 2005: 1.0 initial release.
• 2005: 1.1 adds page joining/splitting, multi-file sequences, deskew speedups, Mac OS X notes, and SPARC Solaris testing.
• 2006: 0.2 adds documentation diagrams and color PPM support.
• 2007: 0.3 moves history to a changelog and adopts conventional tarball layout.
• 2011: 0.4 introduces an Autotools build system and a documented man page.
• 2014: 6 switches file input/output to libav and broadens supported image formats.
• 2022: 7.0.0 migrates the build to Meson and expands ffmpeg 5 compatibility.

Related projects

• unpaper is commonly adjacent to SANE scanner tools, ffmpeg/libav for image decoding and encoding, Netpbm/PNM tooling, ImageMagick conversion, OCR engines, and PDF generation tools.

来源

• https://github.com/unpaper/unpaper/blob/main/AUTHORS
• https://github.com/unpaper/unpaper/blob/main/NEWS
• https://github.com/unpaper/unpaper/blob/main/README.md
• https://github.com/unpaper/unpaper/blob/main/doc/basic-concepts.md
• source_facts.package-manager