OCR Activities

Version: 1.0.0  ·  ID: zoan-packages-ocr  ·  Author: Zoan Software

The OCR (Optical Character Recognition*) package reads text that appears within an image: a scanned PDF, a photo of a document, a screenshot. It’s the add-on to the PDF package: When a document is an image (no selectable text), PDF activities can’t read it — but OCR can.

It uses the Tesseract engine, a widely used open source text recognizer.

OCR is not perfect

OCR interprets an image, so its accuracy depends on the quality of that image: good resolution and sharp text give excellent results; A blurry, crooked, or low-quality scan causes errors. Therefore:

• Use good resolution images (300 DPI or higher for scans).
• Turn on preprocessing (preprocess) to clean up difficult images.
• Validate the extracted data (for example, that a number has the expected format) — do not assume that the text is 100% correct.

Idioma del texto

Almost all activities accept a parameter language with the language code of the text to be recognized:

Value	Idioma
eng	English (default)
spa	Spanish
eng+spa	Several languages at the same time

Indicating the correct language greatly improves precision for accents and locale-specific characters. For Spanish text, always use language = "spa". To see which languages are available, use OCR Get Available Languages.

Image preprocessing

The preprocess parameter applies a cleanup to the image before recognizing it, which improves the result on poor quality scans:

Value	What are you doing
none	No preprocessing (default in most)
auto	Limpieza completa: quita ruido, binariza y endereza
deskew	Just straighten a crooked image
binarize	Convert to black and white (improves contrast)

If recognition fails, try auto

The preprocess = "auto" option solves most problems with skewed or noisy scans. Starts without preprocessing; If the text comes out with errors, upload it to auto.

Two sources: file or screen

• From an image file (.png, .jpg, .bmp, .tiff): OCR Recognize File and structured extraction.
• From the screen: OCR Recognize Screen Region (one area) and OCR Screenshot (the entire screen). These require an interactive desktop session (same as the Desktop package).

Activities

Reconocer texto

Activity	Type	Output	What it does
OCR Recognize File	ocr-recognize	string	Read text from an image file
OCR Recognize Screen Region	ocr-recognize-region	string	Read text from an area of the screen
OCR Screenshot	ocr-screenshot	string	Read all text on the screen

Buscar y ubicar

Activity	Type	Output	What it does
OCR Find Text	ocr-find-text	object	Find the position of a text
OCR Get Lines	ocr-get-lines	List	Structured list of lines and coordinates

Structured extraction

Activity	Type	Output	What it does
OCR Extract Pattern	ocr-extract-pattern	Dictionary	Extract fields with regular expressions
OCR Extract Form Fields	ocr-extract-form	Dictionary	Extrae texto de regiones nombradas
OCR Extract Table	ocr-extract-table	DataTable	Detect a table in the image

Utilidad

Activity	Type	Output	What it does
OCR Get Available Languages	ocr-get-languages	List	Idiomas disponibles

A typical flow

Read a scanned invoice and extract its number:

OCR Recognize File   path = = rutaImagen   language = "spa"   preprocess = "auto"   → output: texto
Set Variable   name = numero   value = = regexFind(texto, "FAC-\d+")

Next steps

• OCR Recognize File — read text from an image.
• OCR Extract Pattern — extract specific fields from a document.
• PDF — for PDF with selectable text (no OCR).