protected override void OnFormClosed(FormClosedEventArgs e)
{
// Clean up
// Save the last setting
Properties.Settings settings = new Properties.Settings();
if (_ocrEngine != null)
{
settings.OcrEngineType = _ocrEngine.EngineType.ToString();
}
settings.Save();
if (_ocrDocument != null)
{
_ocrDocument.Dispose();
_ocrDocument = null;
}
// Dispose the OCR engine (this will call Shutdown as well)
if (_ocrEngine != null)
{
_ocrEngine.Dispose();
_ocrEngine = null;
}
if (_rasterCodecs != null)
{
_rasterCodecs.Dispose();
_rasterCodecs = null;
}
base.OnFormClosed(e);
}
public ProcessDialog(TwainSession twainSession, IOcrEngine ocrEngine, string documentFileName, DocumentFormat format)
{
InitializeComponent();
_twainSession = twainSession;
_document = ocrEngine.DocumentManager.CreateDocument();
_documentFileName = documentFileName;
_format = format;
}
/// <summary>
/// Start a new worker thread for ocr decode
/// </summary>
/// <param name="sImg">Image file path. Only bmp files</param>
/// <param name="sLang">Language source es:IT EN FR..</param>
public String Start(String sFolder, String sImg, String sLang)
{
String sRet="";
try
{
sLang = sLang.ToLower();
_sImgFile = sImg;
if (sLang.ToLower() == "zh")
sLang = "zh-Hant";
_sLang = sLang;
_sFolder = sFolder;
_ocrEngine.LanguageManager.EnableLanguages(new string[] { sLang });
string[] enabledLanguages = _ocrEngine.LanguageManager.GetEnabledLanguages();
Console.WriteLine("Current enabled languages in the engine are:");
foreach (string enabledLanguage in enabledLanguages)
{
// Get the friendly name of this language using the .NET CultureInfo class
CultureInfo ci = new CultureInfo(enabledLanguage);
Console.WriteLine(" {0} ({1})", enabledLanguage, ci.EnglishName);
}
// spell check
IOcrSpellCheckManager spellCheckManager = _ocrEngine.SpellCheckManager;
// Get the spell language supported (languages with a dictionary)
string[] spellLanguages = spellCheckManager.GetSupportedSpellLanguages();
foreach (string spellLanguage in spellLanguages)
Console.WriteLine(spellLanguage);
if (spellCheckManager.IsSpellLanguageSupported(sLang))
{
// Yes, set it
spellCheckManager.SpellLanguage = sLang;
spellCheckManager.Enabled = true;
Console.WriteLine("Current spell language: {0}", spellCheckManager.SpellLanguage);
}
else
spellCheckManager.Enabled = false;
_ocrDocument = _ocrEngine.DocumentManager.CreateDocument();
RasterImage image = _codecs.Load(sImg);
_ocrDocument.Pages.Clear();
_ocrDocument.Pages.AddPage(image, null);
_ocrDocument.Pages.AutoZone(null);
_ocrDocument.Pages.UpdateFillMethod();
sRet = Worker();
}
catch (Exception ex)
{
Program.MainForm.AddLog(String.Format("Exception Worker:{0}", ex.Message), "");
}
return sRet;
}
private void _btnRecognize_Click(object sender, EventArgs e)
{
try
{
using (WaitCursor wait = new WaitCursor())
{
DocumentFormat format = _documentFormatSelector.SelectedFormat;
String documentFileName = _lblImageFileName.Text;
documentFileName = string.Concat(documentFileName, ".", DocumentWriter.GetFormatFileExtension(format));
_ocrPage.Recognize(null);
using (IOcrDocument _document = _ocrEngine.DocumentManager.CreateDocument(null, OcrCreateDocumentOptions.AutoDeleteFile))
{
_document.Pages.Add(_ocrPage);
_document.Save(documentFileName, format, null);
}
// Engine will correct the page deskew before AutoZone or Recognize, so we need to load the page again form the engine.
_viewer.Image = _ocrPage.GetRasterImage();
_viewer.Refresh();
UpdateMyControls();
// if the "View Final Document" option is checked then no need to show this message since
// it will load the saved document file.
if (!_cbViewFinalDocument.Checked)
{
Messager.ShowInformation(this, String.Format("The output document file was saved at ({0})", documentFileName));
}
else
{
if (File.Exists(documentFileName))
{
try
{
Process.Start(documentFileName);
}
catch
{
Messager.ShowError(this, "Unable to open generated results file with external viewing application");
}
}
else
{
Messager.ShowError(this, "Unable to open generated results file with external viewing application.\nThe system cannot find the file specified");
}
}
}
}
catch (Exception ex)
{
Messager.ShowError(this, ex.Message);
}
}
public MainForm()
{
InitializeComponent();
Leadtools.Demos.Support.Unlock();
_ocrEngine = OcrEngineManager.CreateEngine(OcrEngineType.Plus, false);
_ocrEngine.Startup(null, null, null, null);
_ocrDocument = _ocrEngine.DocumentManager.CreateDocument();
}
public MainForm()
{
InitializeComponent();
Leadtools.Demos.Support.Unlock();
_ocrEngine = OcrEngineManager.CreateEngine(OcrEngineType.Plus, false);
_ocrEngine.Startup(null, null, null, null);
_ocrDocument = _ocrEngine.DocumentManager.CreateDocument();
}
private void CleanupImage(RasterImage imageToClean, int startIndex, int count)
{
try
{
//Deskew
if (cleanUpOcrEngine != null && cleanUpOcrEngine.IsStarted)
{
using (IOcrDocument document = cleanUpOcrEngine.DocumentManager.CreateDocument())
{
for (var i = startIndex; i < startIndex + count; i++)
{
imageToClean.Page = i;
document.Pages.AddPage(imageToClean, null);
int angle = -document.Pages[0].GetDeskewAngle();
RotateCommand cmd = new RotateCommand(angle * 10, RotateCommandFlags.Bicubic, RasterColor.FromKnownColor(RasterKnownColor.White));
cmd.Run(imageToClean);
document.Pages.Clear();
}
}
}
else
{
for (var i = startIndex; i < startIndex + count; i++)
{
imageToClean.Page = i;
var deskewCommand = new DeskewCommand();
if (imageToClean.Height > 3500)
{
deskewCommand.Flags = DeskewCommandFlags.DocumentAndPictures | DeskewCommandFlags.DoNotPerformPreProcessing | DeskewCommandFlags.UseNormalDetection | DeskewCommandFlags.DoNotFillExposedArea;
}
else
{
deskewCommand.Flags = DeskewCommandFlags.DeskewImage | DeskewCommandFlags.DoNotFillExposedArea;
}
deskewCommand.Run(imageToClean);
}
}
}
catch (Exception ex)
{
//log.Error(ex);
Console.WriteLine("Not recognized");
}
}
/// <summary>
/// Uses an instance of the LEADTools OCR Engine to "read" the text
/// in the pre-defined OCR region representing the mailing address
/// area. Is thread safe.
/// </summary>
/// <param name="ocrEngine"></param>
/// <param name="document"></param>
/// <returns></returns>
private static string GetAddressBlockText(IOcrEngine ocrEngine, PDFDocument document)
{
string returnedText = null;
using (var codecs = new RasterCodecs())
using (var image = new RasterImage(document.GetPageImage(codecs, 1)))
using (IOcrDocument ocrDocument = ocrEngine.DocumentManager.CreateDocument())
using (IOcrPage ocrPage = ocrDocument.Pages.AddPage(image, null))
{
var myZone = new OcrZone();
myZone.Bounds = new LogicalRectangle(0, 2, 4, 1.4, LogicalUnit.Inch);
ocrPage.Zones.Add(myZone);
ocrPage.Recognize(null);
returnedText = ocrPage.GetText(0).ToUpper();
}
return(returnedText);
}
private bool CleanupImage(RasterImage imageToClean, int startIndex, int count)
{
try
{
if (this.IsStartedOcrEngine())
{
using (IOcrDocument document = this.TheOcrEngine.DocumentManager.CreateDocument())
{
for (var i = startIndex; i < startIndex + count; i++)
{
imageToClean.Page = i;
document.Pages.AddPage(imageToClean, null);
int angle = -document.Pages[0].GetDeskewAngle();
RotateCommand cmd = new RotateCommand(angle * 10, RotateCommandFlags.Bicubic, RasterColor.FromKnownColor(RasterKnownColor.White));
cmd.Run(imageToClean);
document.Pages.Clear();
}
}
}
else
{
for (var i = startIndex; i < startIndex + count; i++)
{
imageToClean.Page = i;
var deskewCommand = new DeskewCommand();
if (imageToClean.Height > 3500)
{
deskewCommand.Flags = DeskewCommandFlags.DocumentAndPictures | DeskewCommandFlags.DoNotPerformPreProcessing | DeskewCommandFlags.UseNormalDetection | DeskewCommandFlags.DoNotFillExposedArea;
}
else
{
deskewCommand.Flags = DeskewCommandFlags.DeskewImage | DeskewCommandFlags.DoNotFillExposedArea;
}
deskewCommand.Run(imageToClean);
}
}
return(true);
}
catch (Exception ex)
{
return(false);
}
}
private void SetDocument(IOcrDocument ocrDocument, string documentFileName)
{
// Delete the old document if it exists
if (_ocrDocument != null)
{
_ocrDocument.Dispose();
}
_lastDocumentFile = documentFileName;
_ocrDocument = ocrDocument;
_ocrPage = _ocrDocument.Pages[0];
BuildWordLists();
_wordTextBox.Text = string.Empty;
SetImage(_ocrPage.GetRasterImage());
UpdateUIState();
}
public static void exe_ocr(string filename, IOcrDocument ocrDocument)
{
Console.WriteLine(filename);
RasterCodecs rasterCodecs = null;
RasterImage rasterImage = null;
try
{
rasterCodecs = new RasterCodecs();
rasterCodecs.ThrowExceptionsOnInvalidImages = false;
rasterImage = rasterCodecs.Load(filename);
AutoBinarizeCommand command = new AutoBinarizeCommand();
command.Run(rasterImage);
//AutoLineRemoveCommand commandLine = new AutoLineRemoveCommand();
//commandLine.Run(rasterImage);
IOcrPage page = ocrDocument.Pages.AddPage(rasterImage, null);
if (page != null)
{
page.UpdateNativeFillMethod();
page.Recognize(null);
}
// ocrDocument.Pages.zon(NativeOcrZoneType.AutoGraphic);
}catch (Exception e)
{
// Console.WriteLine("add image faild: " + e);
}
finally
{
if (rasterCodecs != null)
{
rasterCodecs.Dispose();
}
if (rasterImage != null)
{
rasterImage.Dispose();
}
}
}
public RecognizedWordsDialog(IOcrDocument ocrDocument)
{
InitializeComponent();
_ocrDocument = ocrDocument;
for (int i = 0; i < _ocrDocument.Pages.Count; i++)
{
_pagesListBox.Items.Add(DemosGlobalization.GetResxString(GetType(), "Resx_Page") + (i + 1).ToString());
}
if (_pagesListBox.Items.Count > 0)
{
_pagesListBox.SelectedIndex = 0;
}
// change some of the words listbox in case of Arabic OCR for better words display.
if (_ocrDocument.Engine.EngineType == OcrEngineType.OmniPageArabic)
{
_wordsListBox.RightToLeft = RightToLeft.Yes;
_wordsListBox.Font = new System.Drawing.Font("Times New Roman", 10);
}
}
private void SaveText()
{
if (_imageViewer.HasImage)
{
SaveFileDialog dlg = new SaveFileDialog();
dlg.Filter = "Rich Text File (.rtf)|*.rtf|Text (.txt)|*.txt";
if (dlg.ShowDialog() == DialogResult.OK)
{
string ext = Path.GetExtension(dlg.FileName);
DocumentFormat outputFormat = DocumentFormat.Text;
switch (ext)
{
case ".txt":
outputFormat = DocumentFormat.Text;
break;
case ".rtf":
outputFormat = DocumentFormat.Rtf;
break;
}
using (IOcrDocument document = _ocrEngine.DocumentManager.CreateDocument())
{
document.Pages.AddPage(_imageViewer.Image, null);
document.Pages.Recognize(null);
RtfDocumentOptions rtfOptions = _ocrEngine.DocumentWriterInstance.GetOptions(DocumentFormat.Rtf) as RtfDocumentOptions;
if (rtfOptions != null)
{
rtfOptions.DropObjects = DocumentDropObjects.None;
rtfOptions.TextMode = DocumentTextMode.Framed;
_ocrEngine.DocumentWriterInstance.SetOptions(DocumentFormat.Rtf, rtfOptions);
}
document.Save(dlg.FileName, outputFormat, null);
}
}
}
}
public SaveRecognizedXmlDialog(IOcrDocument ocrDocument)
{
InitializeComponent();
_ocrDocument = ocrDocument;
}
public SaveDocumentDialog(IOcrDocument ocrDocument, DocumentFormat initialFormat, string initialFileName, bool isCustomFileName, string outputDir, bool viewDocument)
{
InitializeComponent();
_ocrDocument = ocrDocument;
_outputDir = outputDir;
// Get the formats
// This is the order of importance, show these first then the rest as they come along
DocumentFormat[] importantFormats =
{
DocumentFormat.Ltd,
DocumentFormat.Pdf,
DocumentFormat.Docx,
DocumentFormat.Rtf,
DocumentFormat.Text,
DocumentFormat.Doc,
DocumentFormat.Xls,
DocumentFormat.Html
};
List <DocumentFormat> formatsToAdd = new List <DocumentFormat>();
Array temp = Enum.GetValues(typeof(DocumentFormat));
List <DocumentFormat> allFormats = new List <DocumentFormat>();
foreach (DocumentFormat format in temp)
{
allFormats.Add(format);
}
// Add important once first:
foreach (DocumentFormat format in importantFormats)
{
formatsToAdd.Add(format);
allFormats.Remove(format);
}
// Add rest
formatsToAdd.AddRange(allFormats);
MyFormat pdfFormat = null;
DocumentWriter docWriter = _ocrDocument.Engine.DocumentWriterInstance;
IOcrDocumentManager ocrDocumentManager = _ocrDocument.Engine.DocumentManager;
string[] engineSupportedFormatNames = ocrDocumentManager.GetSupportedEngineFormats();
foreach (DocumentFormat format in formatsToAdd)
{
bool addFormat = true;
// If this is the "User" or Engines format, only add it if the OCR engine supports them
if (format == DocumentFormat.User && engineSupportedFormatNames.Length == 0)
{
addFormat = false;
}
if (addFormat)
{
string friendlyName;
if (format == DocumentFormat.User)
{
friendlyName = "Engine native";
}
else
{
friendlyName = DocumentWriter.GetFormatFriendlyName(format);
}
MyFormat mf = new MyFormat(format, friendlyName);
_formatComboBox.Items.Add(mf);
if (mf.Format == initialFormat)
{
_formatComboBox.SelectedItem = mf;
}
else if (mf.Format == DocumentFormat.Pdf)
{
pdfFormat = mf;
}
}
switch (format)
{
case DocumentFormat.User:
// Update the User (Engine) options page
{
foreach (string engineFormatName in engineSupportedFormatNames)
{
MyEngineFormat mef = new MyEngineFormat(
engineFormatName,
ocrDocumentManager.GetEngineFormatFriendlyName(engineFormatName));
_userFormatNameComboBox.Items.Add(mef);
if (mef.Format == ocrDocumentManager.EngineFormat)
{
_userFormatNameComboBox.SelectedItem = mef;
}
}
if (_userFormatNameComboBox.SelectedItem == null && _userFormatNameComboBox.Items.Count > 0)
{
_userFormatNameComboBox.SelectedIndex = 0;
}
}
break;
case DocumentFormat.Pdf:
// Update the PDF options page
{
PdfDocumentOptions pdfOptions = docWriter.GetOptions(DocumentFormat.Pdf) as PdfDocumentOptions;
// Clone it in case we change it in the Advance PDF options dialog
_pdfOptions = pdfOptions.Clone() as PdfDocumentOptions;
Array a = Enum.GetValues(typeof(PdfDocumentType));
foreach (PdfDocumentType i in a)
{
// PDFA does NOT support Arabic characters so we are not adding it in case of Arabic OCR engine.
if (i == PdfDocumentType.PdfA && _ocrDocument.Engine.EngineType == OcrEngineType.OmniPageArabic)
{
continue;
}
_pdfDocumentTypeComboBox.Items.Add(i);
}
_pdfDocumentTypeComboBox.SelectedItem = _pdfOptions.DocumentType;
_pdfImageOverTextCheckBox.Checked = _pdfOptions.ImageOverText;
_pdfLinearizedCheckBox.Checked = _pdfOptions.Linearized;
if (string.IsNullOrEmpty(_pdfOptions.Creator))
{
_pdfOptions.Creator = "LEADTOOLS PDFWriter";
}
if (string.IsNullOrEmpty(_pdfOptions.Producer))
{
_pdfOptions.Producer = "LEAD Technologies, Inc.";
}
}
break;
case DocumentFormat.Doc:
// Update the DOC options page
{
DocDocumentOptions docOptions = docWriter.GetOptions(DocumentFormat.Doc) as DocDocumentOptions;
_cbFramedDoc.Checked = (docOptions.TextMode == DocumentTextMode.Framed) ? true : false;
}
break;
case DocumentFormat.Docx:
// Update the DOCX options page
{
DocxDocumentOptions docxOptions = docWriter.GetOptions(DocumentFormat.Docx) as DocxDocumentOptions;
_cbFramedDocX.Checked = (docxOptions.TextMode == DocumentTextMode.Framed) ? true : false;
}
break;
case DocumentFormat.Rtf:
// Update the RTF options page
{
RtfDocumentOptions rtfOptions = docWriter.GetOptions(DocumentFormat.Rtf) as RtfDocumentOptions;
_cbFramedRtf.Checked = (rtfOptions.TextMode == DocumentTextMode.Framed) ? true : false;
}
break;
case DocumentFormat.Html:
// Update the HTML options page
{
HtmlDocumentOptions htmlOptions = docWriter.GetOptions(DocumentFormat.Html) as HtmlDocumentOptions;
Array a = Enum.GetValues(typeof(DocumentFontEmbedMode));
foreach (DocumentFontEmbedMode i in a)
{
_htmlEmbedFontModeComboBox.Items.Add(i);
}
_htmlEmbedFontModeComboBox.SelectedItem = htmlOptions.FontEmbedMode;
_htmlUseBackgroundColorCheckBox.Checked = htmlOptions.UseBackgroundColor;
_htmlBackgroundColorValueLabel.BackColor = MainForm.ConvertColor(htmlOptions.BackgroundColor);
_htmlBackgroundColorLabel.Enabled = _htmlUseBackgroundColorCheckBox.Checked;
_htmlBackgroundColorValueLabel.Enabled = _htmlUseBackgroundColorCheckBox.Checked;
_htmlBackgroundColorButton.Enabled = _htmlUseBackgroundColorCheckBox.Checked;
}
break;
case DocumentFormat.Text:
// Update the TEXT options page
{
TextDocumentOptions textOptions = docWriter.GetOptions(DocumentFormat.Text) as TextDocumentOptions;
Array a = Enum.GetValues(typeof(TextDocumentType));
foreach (TextDocumentType i in a)
{
if (i == TextDocumentType.Ansi)
{
if (textOptions.DocumentType == TextDocumentType.Ansi)
{
textOptions.DocumentType = TextDocumentType.Unicode;
}
if (_ocrDocument.Engine.EngineType == OcrEngineType.OmniPageArabic)
{
continue;
}
}
_textDocumentTypeComboBox.Items.Add(i);
}
_textDocumentTypeComboBox.SelectedItem = textOptions.DocumentType;
_textAddPageNumberCheckBox.Checked = textOptions.AddPageNumber;
_textAddPageBreakCheckBox.Checked = textOptions.AddPageBreak;
_textFormattedCheckBox.Checked = textOptions.Formatted;
}
break;
case DocumentFormat.AltoXml:
// Update the ALTOXML options page
{
AltoXmlDocumentOptions altoXmlOptions = docWriter.GetOptions(DocumentFormat.AltoXml) as AltoXmlDocumentOptions;
_altoXmlFileNameTextBox.Text = altoXmlOptions.FileName;
_altoXmlSoftwareCreatorTextBox.Text = altoXmlOptions.SoftwareCreator;
_altoXmlSoftwareNameTextBox.Text = altoXmlOptions.SoftwareName;
_altoXmlApplicationDescriptionTextBox.Text = altoXmlOptions.ApplicationDescription;
_altoXmlFormattedCheckBox.Checked = altoXmlOptions.Formatted;
_altoXmlIndentationTextBox.Text = altoXmlOptions.Indentation;
_altoXmlSort.Checked = altoXmlOptions.Sort;
_altoXmlPlainText.Checked = altoXmlOptions.PlainText;
_altoXmlShowGlyphInfo.Checked = altoXmlOptions.ShowGlyphInfo;
_altoXmlShowGlyphVariants.Checked = altoXmlOptions.ShowGlyphVariants;
Array a = Enum.GetValues(typeof(AltoXmlMeasurementUnit));
foreach (AltoXmlMeasurementUnit i in a)
{
_altoXmlMeasurementUnit.Items.Add(i);
}
_altoXmlMeasurementUnit.SelectedItem = altoXmlOptions.MeasurementUnit;
}
break;
case DocumentFormat.Ltd:
case DocumentFormat.Emf:
case DocumentFormat.Xls:
case DocumentFormat.Pub:
case DocumentFormat.Mob:
case DocumentFormat.Svg:
default:
// These formats have no options
break;
}
}
// Remove all the tab pages
_optionsTabControl.TabPages.Clear();
// If no format is selected, default to PDF
if (_formatComboBox.SelectedIndex == -1)
{
if (pdfFormat != null)
{
_formatComboBox.SelectedItem = pdfFormat;
}
else
{
_formatComboBox.SelectedIndex = -1;
}
}
_viewDocumentCheckBox.Checked = viewDocument;
_initialFileName = initialFileName;
_isCustomFileName = isCustomFileName;
if (!string.IsNullOrEmpty(_outputDir))
{
MyFormat mf = _formatComboBox.SelectedItem as MyFormat;
char[] trimChars = { '\\' };
_fileNameTextBox.Text = _outputDir.TrimEnd(trimChars) + "\\" + Path.GetFileName(initialFileName);
if (!_isCustomFileName)
{
_fileNameTextBox.Text += "." + GetFileExtension(mf.Format);
}
}
else
{
_fileNameTextBox.Text = initialFileName;
}
_formatComboBox_SelectedIndexChanged(this, EventArgs.Empty);
UpdateUIState();
}
public void RepopulateDocumentInformationControl(IOcrDocument ocrDocument)
{
_lvOcrDocumentInfo.Items[0].SubItems[1].Text = (ocrDocument != null) ? ((ocrDocument.IsInMemory) ? "Memory" : "File") : "None";
_lvOcrDocumentInfo.Items[1].SubItems[1].Text = string.Format("{0}", (ocrDocument != null) ? ocrDocument.Pages.Count : 0);
}
static void Main(string[] args)
{
string licenseFilePath = @"LEADTOOLS.lic";
string developerKey = @"mcxvXsdTqZbnbQrDM9FSk5+RAsBJLhAIot2m3qdpoDO8oK7YMWOw1z6YpXqhCnFE";
RasterSupport.SetLicense(licenseFilePath, developerKey);
// Assuming you added "using Leadtools.Codecs;", "using Leadtools.Forms.Ocr;" and "using Leadtools.Forms.DocumentWriters;" at the beginning of this class
// *** Step 1: Select the engine type and create an instance of the IOcrEngine interface.
// We will use the LEADTOOLS OCR Advantage engine and use it in the same process
IOcrEngine ocrEngine = OcrEngineManager.CreateEngine(OcrEngineType.Professional, false);
// *** Step 2: Startup the engine.
// Use the default parameters
ocrEngine.Startup(null, null, null, @"C:\LEADTOOLS 19\Bin\Common\OcrProfessionalRuntime64");
// *** Step 3: Create an OCR document with one or more pages.
IOcrDocument ocrDocument = ocrEngine.DocumentManager.CreateDocument();
// Add all the pages of a multi-page TIF image to the document
ocrDocument.Pages.AddPages(@"C:\Users\Public\Documents\LEADTOOLS Images\OCR1.tif", 1, -1, null);
// *** Step 4: Establish zones on the page(s), either manually or automatically
// Automatic zoning
ocrDocument.Pages.AutoZone(null);
// *** Step 5: (Optional) Set the active languages to be used by the OCR engine
// Enable English and German languages
ocrEngine.LanguageManager.EnableLanguages(new string[] { "en", "de" });
// *** Step 6: (Optional) Set the spell checking engine
// Enable the spell checking system
ocrEngine.SpellCheckManager.SpellCheckEngine = OcrSpellCheckEngine.Native;
// *** Step 7: (Optional) Set any special recognition module options
// Change the zone method for the first zone in the first page to be Graphics so it will not be recognized
OcrZone ocrZone = ocrDocument.Pages[0].Zones[0];
ocrZone.ZoneType = OcrZoneType.Text;
ocrDocument.Pages[0].Zones[0] = ocrZone;
// *** Step 8: Recognize
ocrDocument.Pages.Recognize(null);
// *** Step 9: Save recognition results
// Save the results to a PDF file
ocrDocument.Save(@"C:\Users\Public\Documents\LEADTOOLS Images\Document.pdf", DocumentFormat.Pdf, null);
ocrDocument.Dispose();
// *** Step 10: Shut down the OCR engine when finished
ocrEngine.Shutdown();
ocrEngine.Dispose();
}
private void DoLoadAndRecognizeDocument(OcrProgressDialog dlg, Dictionary <string, object> args)
{
// Perform load and recognize here
OcrProgressCallback callback = dlg.OcrProgressCallback;
IOcrDocument ocrDocument = null;
try
{
string documentFileName = args["documentFileName"] as string;
ocrDocument = _ocrEngine.DocumentManager.CreateDocument("", OcrCreateDocumentOptions.InMemory);
IOcrPage ocrPage = null;
if (!dlg.IsCanceled)
{
// If we are not using a progress bar, update the description text
if (callback == null)
{
dlg.UpdateDescription("Loading the document (first page only)...");
}
ocrPage = ocrDocument.Pages.AddPage(documentFileName, callback);
}
if (!dlg.IsCanceled)
{
// If we are not using a progress bar, update the description text
if (callback == null)
{
dlg.UpdateDescription("Recognizing the page(s) of the document...");
}
ocrPage.Recognize(callback);
}
if (!dlg.IsCanceled)
{
// We did not cancel, use this document
SetDocument(ocrDocument, documentFileName);
ocrDocument = null;
}
}
catch (Exception ex)
{
ShowError(ex);
}
finally
{
if (callback == null)
{
dlg.EndOperation();
}
// Clean up
if (ocrDocument != null)
{
ocrDocument.Dispose();
}
}
}
private void ThreadProc(object stateInfo)
{
WorkItemData data = (WorkItemData)stateInfo;
IOcrEngine ocrEngine = null;
bool passedCriticalStage = false;
try
{
// See if we have canceled
lock (_abortedLockObject)
{
if (_aborted)
{
return;
}
}
string destinationFile = Path.Combine(data.DestinationDirectory, Path.GetFileName(data.SourceFile));
ocrEngine = data.OcrEngine;
lock (_abortedLockObject)
{
if (_aborted)
{
return;
}
}
// Convert this image file to a document
string extension = DocumentWriter.GetFormatFileExtension(data.Format);
destinationFile = string.Concat(destinationFile, ".", extension);
if (data.Format == DocumentFormat.Ltd && File.Exists(destinationFile))
{
File.Delete(destinationFile);
}
string sourceFile = Path.GetFileName(data.SourceFile);
try
{
// Create a document and add the pages
using (IOcrDocument ocrDocument = ocrEngine.DocumentManager.CreateDocument(null, OcrCreateDocumentOptions.AutoDeleteFile))
{
// Get the image number of pages
int imagePageCount;
RasterCodecs codecs = ocrDocument.RasterCodecsInstance;
using (CodecsImageInfo imageInfo = codecs.GetInformation(data.SourceFile, true))
{
long maximumMemorySize = 42187;
IOcrSettingManager settingManager = ocrEngine.SettingManager;
// Get the maximum size of the bitmap from the setting
if (settingManager.IsSettingNameSupported("Recognition.MaximumPageConventionalMemorySize"))
{
int maximumConventionalMemorySize = settingManager.GetIntegerValue("Recognition.MaximumPageConventionalMemorySize");
maximumMemorySize = (long)maximumConventionalMemorySize * 1024;
}
SetRecommendedLoadingOptions(codecs, imageInfo, maximumMemorySize);
imagePageCount = imageInfo.TotalPages;
}
// Set the DocumentWriter options
using (MemoryStream ms = new MemoryStream(data.DocumentWriterOptions))
{
ocrDocument.DocumentWriterInstance.LoadOptions(ms);
}
passedCriticalStage = true;
//recognize and add pages
for (int pageNumber = 1; pageNumber <= imagePageCount; pageNumber++)
{
lock (_abortedLockObject)
{
if (_aborted)
{
return;
}
}
var image = codecs.Load(data.SourceFile, pageNumber);
using (var ocrPage = ocrEngine.CreatePage(image, OcrImageSharingMode.AutoDispose))
{
ocrPage.Recognize(null);
ocrDocument.Pages.Add(ocrPage);
}
}
// Save
ocrDocument.Save(destinationFile, data.Format, null);
}
}
finally
{
}
OnSuccess(destinationFile);
}
catch (Exception ex)
{
string message;
if (passedCriticalStage && data.FirstTry)
{
message = string.Format("Error '{0}' while converting file '{1}' (first time, quarantined)", ex.Message, data.SourceFile);
AddToQuarantine(data.SourceFile);
}
else if (passedCriticalStage && !data.FirstTry)
{
message = string.Format("Error '{0}' while converting file '{1}' (quarantined error)", ex.Message, data.SourceFile);
}
else
{
message = string.Format("Error '{0}' while converting file '{1}'", ex.Message, data.SourceFile);
}
OnError(message);
}
finally
{
if (ocrEngine != null && ocrEngine != data.OcrEngine)
{
ocrEngine.Dispose();
}
if (Interlocked.Decrement(ref _workItemCount) == 0)
{
_batchFinishedEvent.Set();
}
}
}
public ConvertLdDialog(IOcrDocument ocrDocument, DocumentWriter docWriter, DocumentFormat initialFormat, string initialLdFileName, bool viewDocument)
{
InitializeComponent();
_ocrDocument = ocrDocument;
_docWriter = docWriter;
// Get the formats
// This is the order of importance, show these first then the rest as they come along
DocumentFormat[] importantFormats =
{
DocumentFormat.Pdf,
DocumentFormat.Docx,
DocumentFormat.Rtf,
DocumentFormat.Text,
DocumentFormat.Doc,
DocumentFormat.Xls,
DocumentFormat.Html
};
List <DocumentFormat> formatsToAdd = new List <DocumentFormat>();
Array temp = Enum.GetValues(typeof(DocumentFormat));
List <DocumentFormat> allFormats = new List <DocumentFormat>();
foreach (DocumentFormat format in temp)
{
allFormats.Add(format);
}
// Add important once first:
foreach (DocumentFormat format in importantFormats)
{
formatsToAdd.Add(format);
allFormats.Remove(format);
}
// Add rest
formatsToAdd.AddRange(allFormats);
MyFormat pdfFormat = null;
foreach (DocumentFormat format in formatsToAdd)
{
bool addFormat = true;
// If this is the "User" or Engines format, only add it if the OCR engine supports them
if (format == DocumentFormat.User || format == DocumentFormat.Ltd)
{
addFormat = false;
}
if (addFormat)
{
string friendlyName = DocumentWriter.GetFormatFriendlyName(format);
string extension = DocumentWriter.GetFormatFileExtension(format).ToUpper();
MyFormat mf = new MyFormat(format, friendlyName, extension);
_formatComboBox.Items.Add(mf);
if (mf.Format == initialFormat)
{
_formatComboBox.SelectedItem = mf;
}
else if (mf.Format == DocumentFormat.Pdf)
{
pdfFormat = mf;
}
switch (format)
{
case DocumentFormat.Pdf:
// Update the PDF options page
{
PdfDocumentOptions pdfOptions = docWriter.GetOptions(DocumentFormat.Pdf) as PdfDocumentOptions;
// Clone it in case we change it in the Advance PDF options dialog
_pdfOptions = pdfOptions.Clone() as PdfDocumentOptions;
Array a = Enum.GetValues(typeof(PdfDocumentType));
foreach (PdfDocumentType i in a)
{
_pdfDocumentTypeComboBox.Items.Add(i);
}
_pdfDocumentTypeComboBox.SelectedItem = _pdfOptions.DocumentType;
_pdfImageOverTextCheckBox.Checked = _pdfOptions.ImageOverText;
_pdfLinearizedCheckBox.Checked = _pdfOptions.Linearized;
if (string.IsNullOrEmpty(_pdfOptions.Creator))
{
_pdfOptions.Creator = "LEADTOOLS PDFWriter";
}
if (string.IsNullOrEmpty(_pdfOptions.Producer))
{
_pdfOptions.Producer = "LEAD Technologies, Inc.";
}
}
break;
case DocumentFormat.Doc:
// Update the DOC options page
{
DocDocumentOptions docOptions = docWriter.GetOptions(DocumentFormat.Doc) as DocDocumentOptions;
_cbFramedDoc.Checked = (docOptions.TextMode == DocumentTextMode.Framed) ? true : false;
}
break;
case DocumentFormat.Docx:
// Update the DOCX options page
{
DocxDocumentOptions docxOptions = docWriter.GetOptions(DocumentFormat.Docx) as DocxDocumentOptions;
_cbFramedDocX.Checked = (docxOptions.TextMode == DocumentTextMode.Framed) ? true : false;
}
break;
case DocumentFormat.Rtf:
// Update the RTF options page
{
RtfDocumentOptions rtfOptions = docWriter.GetOptions(DocumentFormat.Rtf) as RtfDocumentOptions;
_cbFramedRtf.Checked = (rtfOptions.TextMode == DocumentTextMode.Framed) ? true : false;
}
break;
case DocumentFormat.Html:
// Update the HTML options page
{
HtmlDocumentOptions htmlOptions = docWriter.GetOptions(DocumentFormat.Html) as HtmlDocumentOptions;
Array a = Enum.GetValues(typeof(DocumentFontEmbedMode));
foreach (DocumentFontEmbedMode i in a)
{
_htmlEmbedFontModeComboBox.Items.Add(i);
}
_htmlEmbedFontModeComboBox.SelectedItem = htmlOptions.FontEmbedMode;
_htmlUseBackgroundColorCheckBox.Checked = htmlOptions.UseBackgroundColor;
_htmlBackgroundColorValueLabel.BackColor = MainForm.ConvertColor(htmlOptions.BackgroundColor);
_htmlBackgroundColorLabel.Enabled = _htmlUseBackgroundColorCheckBox.Checked;
_htmlBackgroundColorValueLabel.Enabled = _htmlUseBackgroundColorCheckBox.Checked;
_htmlBackgroundColorButton.Enabled = _htmlUseBackgroundColorCheckBox.Checked;
}
break;
case DocumentFormat.Text:
// Update the TEXT options page
{
TextDocumentOptions textOptions = docWriter.GetOptions(DocumentFormat.Text) as TextDocumentOptions;
Array a = Enum.GetValues(typeof(TextDocumentType));
foreach (TextDocumentType i in a)
{
_textDocumentTypeComboBox.Items.Add(i);
}
_textDocumentTypeComboBox.SelectedItem = textOptions.DocumentType;
_textAddPageNumberCheckBox.Checked = textOptions.AddPageNumber;
_textAddPageBreakCheckBox.Checked = textOptions.AddPageBreak;
_textFormattedCheckBox.Checked = textOptions.Formatted;
}
break;
case DocumentFormat.AltoXml:
// Update the ALTOXML options page
{
AltoXmlDocumentOptions altoXmlOptions = docWriter.GetOptions(DocumentFormat.AltoXml) as AltoXmlDocumentOptions;
_altoXmlFileNameTextBox.Text = altoXmlOptions.FileName;
_altoXmlSoftwareCreatorTextBox.Text = altoXmlOptions.SoftwareCreator;
_altoXmlSoftwareNameTextBox.Text = altoXmlOptions.SoftwareName;
_altoXmlApplicationDescriptionTextBox.Text = altoXmlOptions.ApplicationDescription;
_altoXmlFormattedCheckBox.Checked = altoXmlOptions.Formatted;
_altoXmlIndentationTextBox.Text = altoXmlOptions.Indentation;
_altoXmlSort.Checked = altoXmlOptions.Sort;
_altoXmlPlainText.Checked = altoXmlOptions.PlainText;
_altoXmlShowGlyphInfo.Checked = altoXmlOptions.ShowGlyphInfo;
_altoXmlShowGlyphVariants.Checked = altoXmlOptions.ShowGlyphVariants;
Array a = Enum.GetValues(typeof(AltoXmlMeasurementUnit));
foreach (AltoXmlMeasurementUnit i in a)
{
_altoXmlMeasurementUnit.Items.Add(i);
}
_altoXmlMeasurementUnit.SelectedItem = altoXmlOptions.MeasurementUnit;
}
break;
case DocumentFormat.Emf:
case DocumentFormat.Xls:
case DocumentFormat.Pub:
case DocumentFormat.Mob:
case DocumentFormat.Svg:
default:
// These formats have no options
break;
}
}
}
// Remove all the tab pages
_optionsTabControl.TabPages.Clear();
// If no format is selected, default to PDF
if (_formatComboBox.SelectedIndex == -1)
{
if (pdfFormat != null)
{
_formatComboBox.SelectedItem = pdfFormat;
}
else
{
_formatComboBox.SelectedIndex = -1;
}
}
_viewDocumentCheckBox.Checked = viewDocument;
_formatComboBox_SelectedIndexChanged(this, EventArgs.Empty);
if (!string.IsNullOrEmpty(initialLdFileName))
{
_ldFileNameTextBox.Text = initialLdFileName;
UpdateOutputFileName();
}
UpdateUIState();
}
