Install ssocr with Homebrew, apt, Nix

Project links

• Homepage
• Repository

Commands and aliases

• ssocr

Project history

The official page explains that ssocr was written because general OCR software struggled with seven-segment displays, often segmenting a single digit into multiple characters. The project's niche is recognizing digits and characters made from seven-segment displays, especially from consistent camera setups.

The first 1.x versions used a separate preprocessing program named ssocrpp. The official version history says version 2 integrated all functionality into one binary and was the first publicly released ssocr version, with later development concentrating on image manipulation options and removing the need for separate tools.

Adoption history

The upstream page lists third-party packages for FreeBSD, NixOS, Debian Buster and newer, Ubuntu Disco Dingo and newer, and Homebrew for macOS and x86-64 GNU/Linux. The supplied package metadata also shows Homebrew, Debian, Nix, and Ubuntu.

Adoption is use-case driven rather than broad platform infrastructure: ssocr is useful for people who need to turn photos of meters, scales, counters, or other seven-segment readouts into scriptable text.

How it is used

ssocr reads an image containing a seven-segment display, applies optional preprocessing commands such as crop, thresholding, inversion, rotation, shear, dilation, and erosion, then prints recognized digits to standard output. The man page documents reading from files or standard input and using Imlib2-supported image formats.

The official usage advice emphasizes fixed camera placement, consistent lighting, debug images, threshold tuning, cropping to the display, deskewing with shear or rotate, and specifying the expected number or character set of digits to catch recognition errors.

Why package nerds care

ssocr is significant because it is tiny, deterministic, and composable: it turns physical-world display readings into stdout, where shell scripts, cron jobs, home automation, and data logging can consume them.

It is also a reminder that not every OCR problem wants Tesseract. For package maintainers, ssocr's value is its small dependency surface, man page, source tarballs, and clear behavior for one specific class of displays.

Timeline

• 2.x: Version 2 integrated ssocrpp preprocessing into the main ssocr binary and became the first publicly released ssocr line.
• 2.8.1: A manual page was added.
• 2.9.0: Reading images from a pipe was added, easing command-line composition.
• 2.11.0: Decimal point detection was added.
• 2.12.0: Hexadecimal digit detection was added.
• 2.13.0: Automatic digit-count determination was added.
• 2.15.0: Minus-sign detection was added.
• 2.22.0: A NEWS file was added, compilation with GCC 10 was fixed, and table-tennis-robot character sets were added.
• 2025: The official page lists version 2.25.1 as the current source tarball and generated man page version.

Related projects

• The official page discusses general OCR programs such as GNU Ocrad, GOCR, OCRFeeder, ocropus, and Tesseract as related but less suitable for this specific seven-segment problem.
• It also links related seven-segment OCR projects such as Display OCR and sevenSegDecode.

Sources

• https://www.unix-ag.uni-kl.de/~auerswal/ssocr
• https://www.unix-ag.uni-kl.de/~auerswal/ssocr/ssocr-manpage.html
• input source_facts.package-manager