//called when data for any output pin is requested
public void Evaluate(int SpreadMax)
{
if (_ocr == null)
{
try
{
_ocr = new Tesseract(@FTessdata[0], "eng", Tesseract.OcrEngineMode.OEM_TESSERACT_ONLY);
}
catch (Exception exception){
FLogger.Log(LogType.Debug, exception.Message);
}
}
if (FInit.IsChanged && FInit[0] == true)
{
Image<Bgr, byte> My_Image = new Image<Bgr, byte>(@FInput[0]);
Image<Gray, byte> gray = My_Image.Convert<Gray, Byte>();
_ocr.Recognize(gray);
FOutput[0] = _ocr.GetText();
}
//FOutput[0] = "hallo";
}
private void btnAnalyzeImage_Click(object sender, EventArgs e)
{
if(picBox.Image == null)
{
MessageBox.Show("Load an image first!");
}
else
{
statusLabelOCR.Text = "Analyzing invoice image...";
Task.Run(() =>
{
using (var img = new Image<Bgr, byte>(path))
{
string tessdata = Environment.GetEnvironmentVariable("EMGU_ROOT") + @"\bin\tessdata";
using (var ocrProvider = new Tesseract(tessdata, "eng", OcrEngineMode.TesseractCubeCombined))
{
ocrProvider.Recognize(img);
string text = ocrProvider.GetText().TrimEnd();
rtbOcrResult.Invoke((MethodInvoker)delegate
{
statusLabelOCR.Text = "Analysis completed.";
rtbOcrResult.AppendText(text);
});
}
}
});
}
}
/// <summary>
/// Initialize a new OCR Object.
///
/// This object is a wrapper for the Emgu Tesseract Wrapper to give a level of abstraction
/// necessary for scanning shreds
/// </summary>
/// <param name="accuracy"> Desired Accuracy setting </param>
/// <param name="language"> Language of text on image used for OCR model </param>
/// <param name="enableTimer"> Set enable Timer to true to measure scan time for diagnostic purposes </param>
public OCR(Accuracy accuracy = Accuracy.High, string language = "eng", bool enableTimer = false)
{
_timer = new Stopwatch();
if (enableTimer)
{
_timer.Start();
}
Tesseract.OcrEngineMode mode = Tesseract.OcrEngineMode.OEM_TESSERACT_CUBE_COMBINED;
switch (accuracy)
{
case Accuracy.Low:
mode = Tesseract.OcrEngineMode.OEM_TESSERACT_ONLY;
break;
case Accuracy.Medium:
mode = Tesseract.OcrEngineMode.OEM_CUBE_ONLY;
break;
case Accuracy.High:
mode = Tesseract.OcrEngineMode.OEM_TESSERACT_CUBE_COMBINED;
break;
}
_tesseract = new Tesseract("tessdata", language, mode);
//"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz");
_tesseract.SetVariable("tessedit_unrej_any_wd", "T");
//_tesseract.SetVariable("tessedit_char_whitelist","abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWZ0123456789.,");
_text = null;
_chars = null;
_confidence = -1;
}
private void InitOcr(String path, String lang, OcrEngineMode mode)
{
try
{
if (_ocr != null)
{
_ocr.Dispose();
_ocr = null;
}
if (String.IsNullOrEmpty(path))
{
path = ".";
}
TesseractDownloadLangFile(path, lang);
TesseractDownloadLangFile(path, "osd"); //script orientation detection
String pathFinal = path.Length == 0 ||
path.Substring(path.Length - 1, 1).Equals(Path.DirectorySeparatorChar.ToString())
? path
: String.Format("{0}{1}", path, System.IO.Path.DirectorySeparatorChar);
_ocr = new Emgu.CV.OCR.Tesseract(pathFinal, lang, mode);
}
catch (System.Net.WebException e)
{
_ocr = null;
throw new Exception("Unable to download tesseract lang file. Please check internet connection.", e);
}
catch (Exception e)
{
_ocr = null;
}
}
public EmguOcr()
{
_tesseract = new Emgu.CV.OCR.Tesseract
(
Environments.AppConfig.TesseractDataPath,
Environments.AppConfig.OcrLanguage,
OcrEngineMode.Default);
}
public Form1()
{
InitializeComponent();
ocr = new Tesseract();
ocr.SetVariable("tessedit_char_whitelist", "1234567890");
ocr.Init(@"D:\", "eng", OcrEngineMode.TesseractOnly);
}
static void Main(string[] args)
{
using (var image = new Image<Bgr, byte>(Path.GetFullPath("text.png")))
{
using (var tess = new Tesseract("", "eng", OcrEngineMode.TesseractCubeCombined))
{
tess.Recognize(image);
var text = tess.GetText().TrimEnd();
}
}
}
/// <summary>
/// Initializes and creates the tessearct engine
/// </summary>
public WrapperTesseract(string path, string lng)
{
// Variable value is set to parent directory of tessdata
//string path = AppDomain.CurrentDomain.BaseDirectory;
//Path should be same as TESSDATA folder
//Environment.SetEnvironmentVariable("TESSDATA_PREFIX", path); //This path should always end with a "/" or "\", e.g., TESSDATA_PREFIX="/usr/share/tesseract-ocr/"
_ocr = new Tesseract(path, lng, Tesseract.OcrEngineMode.OEM_TESSERACT_CUBE_COMBINED);
///_ocr = new Tesseract(path, "eng", Tesseract.OcrEngineMode.OEM_TESSERACT_CUBE_COMBINED);
}
public static void test()
{
Tesseract _ocr = new Tesseract(@"C:\Users\yaoyichi\Desktop\emgucv-windows-universal-cuda 2.9.0.1922\", "eng", Tesseract.OcrEngineMode.OEM_TESSERACT_CUBE_COMBINED);
string dir = @"C:\Users\yaoyichi\Desktop\";
string fn = "100890509_mc1024_k64.png";
Image<Bgr, Byte> gray = new Image<Bgr, byte>(dir + fn);
_ocr.Recognize(gray);
//TextRecognitionWorker trw = new TextRecognitionWorker();
//trw.ApplyABBYYOCR(dir, fn);
}
private void TextRecogniser(Image image)
{
Emgu.CV.OCR.Tesseract tesseract = new Emgu.CV.OCR.Tesseract
(@"F:\Working Pack\c#\Scripter\Scripter\Templates\eng.traineddata", "eng",
OcrEngineMode.TesseractLstmCombined);
var tempImg = new Bitmap(image).ToImage <Gray, byte>();
tesseract.SetImage(tempImg);
tesseract.Recognize();
string response = tesseract.GetUTF8Text();
tesseract.Dispose();
}
public static Tesseract.Charactor[] DoOcr(Image<Gray, byte> img)
{
Tesseract.Charactor[] chars;
string text;
using (var tesseract = new Tesseract("tessdata", "eng", Tesseract.OcrEngineMode.OEM_TESSERACT_CUBE_COMBINED)
)
{
tesseract.Recognize(img);
text = tesseract.GetText();
chars = tesseract.GetCharactors();
}
Console.WriteLine(text);
Console.ReadLine();
return chars;
}
/// <summary>
/// Calculates the cost by summing the unique cost of each word
/// </summary>
/// <param name="chars"> Tesseract OCR Charactor results </param>
/// <returns> Cost, where zero is perfect and long.MaxValue is worst </returns>
public static long Cost(Tesseract.Charactor[] chars)
{
if (chars.Length < 1)
{
return 0;
}
long total = 0;
double current = -1;
foreach (Tesseract.Charactor charactor in chars)
{
if (Math.Abs(current - charactor.Cost) > 0.0001)
{
current = charactor.Cost;
total += (long)current;
}
}
return total;
}
private void InitOcr(String path, String lang, OcrEngineMode mode)
{
try
{
if (_ocr != null)
{
_ocr.Dispose();
_ocr = null;
}
_ocr = new Tesseract(path, lang, mode);
languageNameLabel.Text = String.Format("{0} : {1}", lang, mode.ToString());
}
catch (Exception e)
{
_ocr = null;
MessageBox.Show(e.Message, "Failed to initialize tesseract OCR engine", MessageBoxButtons.OK);
languageNameLabel.Text = "Failed to initialize tesseract OCR engine";
}
}
public BotCV(IntPtr win, int Hmin, int Smin, int Vmin, int Hmax, int Smax, int Vmax)
{
this.Hmin = Hmin;
this.Hmax = Hmax;
this.Smin = Smin;
this.Smax = Smax;
this.Vmin = Vmin;
this.Vmax = Vmax;
sc = new ScreenCapture();
winDiscript = win;
try
{
tess = new Emgu.CV.OCR.Tesseract(@"", "eng", OcrEngineMode.TesseractOnly);
tess.SetVariable("tessedit_char_whitelist", ".0123456789");
}
catch (Exception te)
{
Console.WriteLine(te.Message);
tess = null;
}
}
private void button5_Click(object sender, EventArgs e)
{
var _ocr = new Tesseract(@"D:\", "eng", OcrEngineMode.TesseractCubeCombined);
_ocr.SetVariable("tessedit_char_whitelist", "1234567890");
_ocr.Recognize(_inputArray);
var result = _ocr.GetCharacters().ToList();
var sb = new StringBuilder();
result.ForEach(c => sb.Append(c.Text));
label1.Text = sb.ToString();
}
public Form1()
{
InitializeComponent();
_ocr = new Tesseract("", "eng", Tesseract.OcrEngineMode.OEM_TESSERACT_CUBE_COMBINED);
}
/// <summary>
/// Create a license plate detector
/// </summary>
public LicensePlateDetector()
{
//create OCR engine
_ocr = new Tesseract("tessdata", "eng", Tesseract.OcrEngineMode.OEM_TESSERACT_CUBE_COMBINED);
_ocr.SetVariable("tessedit_char_whitelist", "ABCDEFGHIJKLMNOPQRSTUVWXYZ-1234567890");
}
//główne operacje na obrazie
public void ocr()
{
//otworzenie pliku
FileStream srcstream = new FileStream(pic_file, FileMode.Open, FileAccess.Read, FileShare.ReadWrite);
//stworzenie bitmapy
Bitmap source = new Bitmap(srcstream);
//zmiana ustawień webform
Panel1.Visible = false;
Image1.Dispose();
Label2.Text = "Processing...";
Panel3.Visible = true;
//Preperation code
Bitmap ext = source;
//AForge.Imaging.Filters.
//Przekształcenie obrazu na skalę odcieni szarości - testować dla obrazów o różnej kolorystyce(opracować system wyznaczania parametrów filtru na podstawie RGB zdjęcia)
AForge.Imaging.Filters.Grayscale grScl = new AForge.Imaging.Filters.Grayscale(0.2125, 0.0154, 0.0721 );
source = grScl.Apply(source);
//Zwiększenie kontrastu
AForge.Imaging.Filters.ContrastStretch conCor = new AForge.Imaging.Filters.ContrastStretch();
source = conCor.Apply(source);
//Wyostrzenie
AForge.Imaging.Filters.Sharpen shp = new AForge.Imaging.Filters.Sharpen();
source = shp.Apply(source);
//Segmentation code
bool procesed = false;
// Image2.Width = 350;
// Image2.Height = (int)((source.Height * 200) / source.Width);
try
{
Emgu.CV.Image<Bgr, Byte> to_rec = new Emgu.CV.Image<Bgr, byte>(source);
Do_ocr = new Tesseract("tessdata", "eng", Tesseract.OcrEngineMode.OEM_DEFAULT);
try
{
Do_ocr.Recognize<Bgr>(to_rec);
//recognizedText.Text = ocr.GetText();
PastOCRBox.Text = Do_ocr.GetText();
// StatusBox.Text = "Finished! Ready for next one...";
Do_ocr.Dispose();
to_rec.Dispose();
}
catch (Exception exp)
{
Label2.Text = "Recognition error! " + exp.Message;
Do_ocr.Dispose();
return;
}
}
catch (Exception exp)
{
Label2.Text = "OCR engine failed! " + exp.Message;
return;
}
//czyszczenie z plików tymczasowych
// source.Save("D:\\test.bmp");
// ext.Save("D:\\testcor.bmp");
source.Dispose();
srcstream.Close();
srcstream.Dispose();
//System.IO.File.Delete(pic_file);
System.IO.File.Delete(Server.MapPath("~/img/prev.bmp"));
System.IO.File.Delete(Server.MapPath("~/img/tmp.bmp"));
//przygotować wygląd strony po rozpoznawaniu
Panel3.Visible = false;
Label1.Visible = false;
Panel0.Visible = false;
Panel5.Visible = false;
Panel4.Visible = true;
}
public ScreenAnalyzer(string languageCode)
{
this.language = Language.codeToLanguage(languageCode);
ocr = new Tesseract("Assets\\", languageCode, language.OcrMode);
//ocr.SetVariable("tessedit_char_whitelist", language.OcrWhitelist);
}
private void server()
{
int port = 8001;
IPAddress ipAd = IPAddress.Parse(ip);
// use local m/c IP address, and
// use the same in the client
ASCIIEncoding asen = new ASCIIEncoding();
/* Initializes the Listener */
TcpListener myList = new TcpListener(IPAddress.Any, port);
Console.WriteLine("The server is running at port 8001...");
Console.WriteLine("The local End point is :" +
myList.LocalEndpoint);
/* Start Listeneting at the specified port */
byte[] b = new byte[2000000];
myList.Start();
while (!stoped)
{
char[] recived = new char[2000000];
Console.WriteLine("Waiting for a connection.....");
TcpClient s = myList.AcceptTcpClient();
Console.WriteLine("Connection accepted " +s.Client.RemoteEndPoint);
NetworkStream stream = s.GetStream();
var reader = new StreamReader(stream);
var k = stream.Read(b, 0, 2000000);
Console.WriteLine("Recieved...");
ImageConverter conv = new ImageConverter();
Bitmap rec = new Bitmap((Bitmap)conv.ConvertFrom(b));
Console.WriteLine("Recognicion...");
Tesseract ocr = new Tesseract("tessdata", "pol+eng", Tesseract.OcrEngineMode.OEM_DEFAULT);
ocr.Recognize<Bgr>(new Emgu.CV.Image<Bgr, Byte>(rec));
string to_send_str = ocr.GetText();
ocr.Dispose();
Console.Write(to_send_str + "\n");
stream.Write(asen.GetBytes(to_send_str),0,to_send_str.Length);
Console.WriteLine("\nSent Acknowledgement");
/* clean up */
s.Close();
}
myList.Stop();
Console.WriteLine("Server stopped");
}
static void Main(string[] args)
{
var files = Directory.GetFiles(@"d:\raw\classified\");
//var img = new Image<Rgb, byte>(files[0]);
var img = new Image<Emgu.CV.Structure.Gray, byte>(@"d:\raw\classified\test.jpg");
var _ocr = new Tesseract(@"D:\", "eng", OcrEngineMode.TesseractCubeCombined);
_ocr.SetVariable("tessedit_char_whitelist", "1234567890");
_ocr.Recognize(img);
var result = _ocr.GetCharacters();
}
public FaceDetection()
{
InitializeComponent();
_ocr = new Tesseract();
}
private static void Setup()
{
if (!_setup)
{
_ocr = new Tesseract("", "eng", Tesseract.OcrEngineMode.OEM_TESSERACT_CUBE_COMBINED);
// If this is not done cost for any letters recognised which are not letters (i.e. false positives) will be zero
_ocr.Recognize(new Image<Gray, Byte>("tessdata/OCR initialise data.jpg"));
for (int i = 0; i < _mapHexColours.Length; i++ )
{
_mapColours.Add(ColorTranslator.FromHtml((_mapHexColours[i])));
}
_setup = true;
}
}
/// <summary>
/// Create a license plate detector
/// </summary>
/// <param name="dataPath">
/// The datapath must be the name of the parent directory of tessdata and
/// must end in / . Any name after the last / will be stripped.
/// </param>
public LicensePlateDetector(String dataPath)
{
//create OCR engine
_ocr = new Tesseract(dataPath, "eng", OcrEngineMode.TesseractCubeCombined);
_ocr.SetVariable("tessedit_char_whitelist", "ABCDEFGHIJKLMNOPQRSTUVWXYZ-1234567890");
}
//private TesseractEngine engine;
public TesseractManager()
{
engineJpn = new Emgu.CV.OCR.Tesseract(@"./tessdata", "jpn", OcrEngineMode.LstmOnly);
engineUma = new Emgu.CV.OCR.Tesseract(@"./tessdata", "uma", OcrEngineMode.LstmOnly);
//engine = new TesseractEngine(@"./tessdata", "jpn", EngineMode.Default);
}