Link Search Menu Expand Document

Use Image Preprocessing Filters - PowerShell

Text Recognition SDK sample in PowerShell demonstrating ‘Use Image Preprocessing Filters’

# Add reference to ByteScout.TextRecognition.dll assembly
Add-Type -Path "c:\Program Files\ByteScout Text Recognition SDK\net40\ByteScout.TextRecognition.dll"

$InputDocument = "skewed.png"
$OutputDocument = ".\result.txt"

# Create and activate TextRecognizer instance
$textRecognizer = New-Object ByteScout.TextRecognition.TextRecognizer
$textRecognizer.RegistrationName = "demo"
$textRecognizer.RegistrationKey = "demo"

try {
    # Load document (image or PDF)

    # Set the location of OCR language data files
    $textRecognizer.OCRLanguageDataFolder = "c:\Program Files\ByteScout Text Recognition SDK\ocrdata_best\"

    # Set OCR language.
    # "eng" for english, "deu" for German, "fra" for French, "spa" for Spanish, etc. - according to files in "ocrdata" folder
    # Find more language files at
    $textRecognizer.OCRLanguage = "eng"

    # Add deskew filter that automatically rotates the image to make the text horizontal.
    # Note, it analyzes the left edge of scanned text. Any dark artifacts may prevent 
    # the correct angle detection.

    # Other filters that may be useful to improve recognition
    # (note, the filters are applied in the order they were added):

    # Improve image contrast.

    # Apply gamma correction.

    # Apply median filter. Helps to remove noise.

    # Apply dilate filter. Helps to cure symbols erosion.

    # Lines removers. Removing borders of some tables may improve the recognition.

    # Recognize text from all pages and save it to file

    # Open the result file in default associated application (for demo purposes)
    & $OutputDocument
catch {
    # Display exception
    Write-Host $_.Exception.Message


@echo off

powershell -NoProfile -ExecutionPolicy Bypass -Command "& .\ImagePreprocessingFilters.ps1"
echo Script finished with errorlevel=%errorlevel%


Download Source Code (.zip)

Return to the previous page Explore Text Recognition SDK