private void doScan(string targetFolderName, string resultFileName)
{
var bitmapConverter = new Tesseract.BitmapToPixConverter();
var tesseract = new Tesseract.TesseractEngine(@"C:\Program Files (x86)\Tesseract-OCR\tessdata", "jpn", @"digits");
//tesseract.SetVariable("tessedit_char_whitelist", "0123456789");
tesseract.DefaultPageSegMode = Tesseract.PageSegMode.SingleChar;
MessageBox.Show(tesseract.Version);
var fileNames = Directory.GetFiles(targetFolderName, "*.jpg");
Array.ForEach(fileNames, fileName => {
var originalBitmap = new System.Drawing.Bitmap(fileName);
//MessageBox.Show($"{ originalBitmap.Width }*{ originalBitmap.Height }");
var targetRect = new System.Drawing.Rectangle(2980, 140, 3700 - 2980, 330 - 140);
/// var targetRect = new System.Drawing.Rectangle(1500, 75, 1875 - 1500, 170 - 75);
var targetBitmap = originalBitmap.Clone(targetRect, System.Drawing.Imaging.PixelFormat.DontCare);
using (var page = tesseract.Process(bitmapConverter.Convert(targetBitmap)))
{
MessageBox.Show(page.GetText());
}
var targetRect2 = new System.Drawing.Rectangle(1390, 1370, 2390 - 1390, 1530 - 1370);
/// var targetRect2 = new System.Drawing.Rectangle(720, 690, 1230 - 720, 800 - 690);
var targetBitmap2 = originalBitmap.Clone(targetRect2, System.Drawing.Imaging.PixelFormat.DontCare);
using (var page = tesseract.Process(bitmapConverter.Convert(targetBitmap2)))
{
MessageBox.Show(page.GetText());
}
});
}
public static string OCRImageWithTesseract(Bitmap theBmp)
{
var ocrText = string.Empty;
try
{
using (var engine = new Tesseract.TesseractEngine(@".\tessdata\", "eng", Tesseract.EngineMode.LstmOnly))
{
var pix = Tesseract.PixConverter.ToPix(theBmp);
using (var tessPage = engine.Process(pix))
{
ocrText = tessPage.GetText();
}
using (var tessPage = engine.Process(pix, Tesseract.PageSegMode.Auto))
{
ocrText = tessPage.GetText().Trim();
}
}
}
catch (Exception ex)
{
if (ex is AccessViolationException || ex is InvalidOperationException)
{
Trace.TraceError("Tesseract Error: " + ex.Message);
}
else
{
throw;
}
}
return(ocrText);
}
private void DoOCR()
{
string s_people = "",
s_mail = "";
try
{
var bitmap = screen_rgb.ToBitmap();
using (var engine = new Tesseract.TesseractEngine(@"./tessdata", "eng", Tesseract.EngineMode.Default))
{
engine.SetVariable("tessedit_char_whitelist", "0123456789");
var people = bitmap.Clone(new Rectangle(230, 250, 130, 25), System.Drawing.Imaging.PixelFormat.Format32bppArgb);
using (var page = engine.Process(people))
{
s_people = Regex.Replace(page.GetText(), "[^0-9]", "");
}
var mail = bitmap.Clone(new Rectangle(395, 250, 182, 25), System.Drawing.Imaging.PixelFormat.Format32bppArgb);
using (var page = engine.Process(mail))
{
s_mail = Regex.Replace(page.GetText(), "[^0-9]", "");
}
}
if (string.IsNullOrEmpty(s_people))
{
s_people = "0";
}
if (string.IsNullOrEmpty(s_mail))
{
s_mail = "0";
}
if (s_mail != "0" && s_people != "0")
{
Stats.Add(DateTime.Now, new Tuple <int, int>(int.Parse(s_people), int.Parse(s_mail)));
}
if (!IsActive)
{
return;
}
Invoke(new Action(() => { PictureBoxOutput.Image = bitmap; }));
}
catch (Exception e)
{
AddOutput("OCR Error [$:'" + s_people + "', M:'" + s_mail + "'] " + e.Message + e.StackTrace);
}
}
/// <summary>
/// 光学文字認識による画像から文字列の読み出し
/// </summary>
/// <param name="bmp">元の画像</param>
/// <param name="targetLanguageISO639_2">抽出する言語コード</param>
/// <returns>抽出した文字列</returns>
public static string ToOcrString(this Bitmap bmp, string targetLanguageISO639_2 = "eng")
{
using (var tesseract = new Tesseract.TesseractEngine(DataProperty.TessDataFolder, targetLanguageISO639_2))
{
return(tesseract.Process(bmp).GetText().Trim());
}
}
public static OCRResult OCRImageWithTesseract(Bitmap theBmp)
{
var ocrResult = new OCRResult()
{
OcrText = string.Empty, HocrText = string.Empty, MeanConfidence = 0.0f
};
try
{
using (var engine = new Tesseract.TesseractEngine(@".\tessdata\", "eng", Tesseract.EngineMode.LstmOnly))
{
using (var pix = Tesseract.PixConverter.ToPix(theBmp))
{
using (var tessPage = engine.Process(pix, Tesseract.PageSegMode.Auto))
{
ocrResult.OcrText = tessPage.GetText();
ocrResult.HocrText = tessPage.GetHOCRText(1);
ocrResult.MeanConfidence = tessPage.GetMeanConfidence();
}
}
}
}
catch (Exception ex)
{
if (ex is AccessViolationException || ex is InvalidOperationException)
{
Trace.TraceError("Tesseract Error: " + ex.Message);
}
else
{
throw;
}
}
return(ocrResult);
}
/***********************************************************************************************************************
*
* Tesseract-OCR
*
************************************************************************************************************************/
/// <summary>
///
/// </summary>
/// <param name="sender"></param>
/// <param name="e"></param>
/// <example>
/// http://www.kanenote.org/blog/?p=1228
/// https://qiita.com/nobi1234/items/c8d7b748c1aa31d771a1
/// http://whoopsidaisies.hatenablog.com/entry/2013/12/16/174819
/// https://tesseract-ocr.github.io/tessdoc/Data-Files.html
/// https://kakusuke98.hatenablog.com/entry/2019/11/14/004609
/// http://blog.qes.co.jp/2017/06/
///
/// </example>
private void btn_tesseract_ocr_Click(object sender, EventArgs e)
{
//言語ファイルの格納先
//言語（日本語なら"jpn"）
//string langPath = @"C:\tessdata";
string langPath = strFolderPath + @"\tessdata";
string lngStr = "eng";
//画像ファイル
//var img = new Bitmap(@"C:\Temp\test.jpg");
Bitmap img = (Bitmap)picBoxDst.Image;
// OCRの実行
//OpenCvSharp.Text.OCRTesseract.Create("");
using (var tesseract = new Tesseract.TesseractEngine(langPath, lngStr)) {
Tesseract.Page page = tesseract.Process(img); // OCRの実行
//System.Console.Write(page.GetText()); //表示
MessageBox.Show(page.GetText());
//Console.WriteLine(page.GetText());
//Console.ReadLine();
}
}
public static void processImage(Tesseract.TesseractEngine engine, Tesseract.Pix img, string testImagePath, int pageNum)
{
System.IO.FileInfo mfile = new System.IO.FileInfo(testImagePath);
string hocrPath = mfile.DirectoryName + "\\" + System.IO.Path.GetFileNameWithoutExtension(testImagePath) + "_p" + pageNum.ToString() + ".xhtml";
using (var page = engine.Process(img, PageSegMode.AutoOsd))
{
page.AnalyseLayout();
string hocr = page.GetHOCRText(0, true);
System.IO.File.AppendAllText(hocrPath, hocr);
var text = page.GetText();
Console.WriteLine("Mean confidence: {0}", page.GetMeanConfidence());
Console.WriteLine("Text (GetText): \r\n{0}", text);
Console.WriteLine("Text (iterator):");
using (var iter = page.GetIterator())
{
iter.Begin();
do
{
do
{
do
{
do
{
if (iter.IsAtBeginningOf(PageIteratorLevel.Block))
{
Console.WriteLine("<BLOCK>");
Rect currentBlock;
iter.TryGetBoundingBox(PageIteratorLevel.Block, out currentBlock);
Console.WriteLine(iter.BlockType.ToString());
Console.WriteLine("(" + currentBlock.X1.ToString() + "," + currentBlock.Y1.ToString() + ") (" + currentBlock.X2.ToString() + "," + currentBlock.Y2.ToString() + ")");
Console.WriteLine("");
}
Console.Write(iter.GetText(PageIteratorLevel.Word));
Console.Write(" ");
if (iter.IsAtFinalOf(PageIteratorLevel.TextLine, PageIteratorLevel.Word))
{
Console.WriteLine(iter.BlockType.ToString());
}
} while (iter.Next(PageIteratorLevel.TextLine, PageIteratorLevel.Word));
if (iter.IsAtFinalOf(PageIteratorLevel.Para, PageIteratorLevel.TextLine))
{
Console.WriteLine();
}
} while (iter.Next(PageIteratorLevel.Para, PageIteratorLevel.TextLine));
} while (iter.Next(PageIteratorLevel.Block, PageIteratorLevel.Para));
} while (iter.Next(PageIteratorLevel.Block));
}
}
}
private void Button_Click(object sender, RoutedEventArgs e)
{
int err_count = 0;
while (err_count < 2)
{
Microsoft.VisualBasic.Interaction.AppActivate("Overwatch");
SendKeys.SendWait("{PRTSC}");
System.Diagnostics.Process[] processList = System.Diagnostics.Process.GetProcesses();
if (System.Windows.Clipboard.ContainsImage())
{
string lngStr = System.IO.Path.Combine(System.IO.Directory.GetCurrentDirectory(), @"tesseract-ocr\\");
string langPath = "eng";
var img = (Bitmap)System.Windows.Forms.Clipboard.GetImage();
var bitmap24 = new Bitmap(img.Width, img.Height, System.Drawing.Imaging.PixelFormat.Format24bppRgb);
using (var gr = Graphics.FromImage(bitmap24))
{
gr.DrawImage(img, new Rectangle(0, 0, bitmap24.Width, bitmap24.Height));
}
using (var tesseract = new Tesseract.TesseractEngine(lngStr, langPath))
{
tesseract.SetVariable("SEARCH", "1");
Tesseract.Page page = tesseract.Process(bitmap24);
if (page.GetText().Contains("SEARCH") || page.GetText().Contains("TIME") || page.GetText().Contains("ELAPSED") || page.GetText().Contains("LEAGUE") || page.GetText().Contains("HIGHLIGHTS") || page.GetText().Contains("SOCIAL") || page.GetText().Contains("EXIT"))
{
err_count = 0;
}
else
{
err_count += 1;
}
}
}
Thread.Sleep(50000);
}
DateTime dt = DateTime.Now;
string result = dt.ToString("yyyy/MM/dd HH:mm:ss");
string lnk = "https://maker.ifttt.com/trigger/match/with/key/xxxxxxxx/?value1=" + result;
System.Diagnostics.Process.Start(lnk);
Environment.Exit(0);
}
}//311行
//●文字認識処理
//・引数 Bitmap img ：文字認識処理対象の画像を指定する
// string lang：文字認識処理を行う言語を指定する
//・戻り値：文字認識処理結果
private string chara_recog(Bitmap img, string lang)
{
//文字認識結果を格納する変数
string str;
// OCRを行うオブジェクトの生成
// 言語データの場所と言語名を引数で指定する
var tesseract = new Tesseract.TesseractEngine(
@"..\..\..\tessdata", // 言語ファイルを「C:\tessdata」に置いた場合
lang); // 英語なら"eng" 「○○.traineddata」の○○の部分
// OCRの実行と表示
var page = tesseract.Process(img);
str = page.GetText();
//文字認識結果を返す
return(str);
}//311行
private void Button1_Click(object sender, EventArgs e)
{
using (var bmp = Snip())
{
if (bmp != null)
{
using (var image = new Bitmap(bmp))
{
using (var ocr = new Tesseract.TesseractEngine(@"./tessdata", "eng", Tesseract.EngineMode.Default))
{
ocr.SetVariable("tessedit_char_whitelist", "0123456789");
using (var result = ocr.Process(image, Tesseract.Rect.Empty))
{
Clipboard.SetText(result.GetText().Trim());
}
}
}
}
}
}
public static T4Result OCRImageResult(Bitmap theBmp, bool getHOCR = false, bool getXHTML = false, bool getJSON = false)
{
var result = new T4Result();
try
{
using (var engine = new Tesseract.TesseractEngine(@".\tessdata\", "eng", Tesseract.EngineMode.LstmOnly))
{
var pix = Tesseract.PixConverter.ToPix(theBmp);
using (var tessPage = engine.Process(pix, Tesseract.PageSegMode.Auto))
{
result.MeanConfidence = tessPage.GetMeanConfidence();
result.Text = tessPage.GetText();
if (getHOCR)
{
result.HOCR = tessPage.GetHOCRText(1);
}
if (getXHTML)
{
result.XHTML = tessPage.GetHOCRText(1, true);
}
if (getJSON)
{
result.JSON = GetJSON(tessPage);
}
}
}
}
catch (Exception ex)
{
if (ex is AccessViolationException || ex is InvalidOperationException)
{
Trace.TraceError("Tesseract Error: " + ex.Message);
}
else
{
throw;
}
}
return(result);
}
public void doOcr()
{
Bitmap resizer = new Bitmap(toOcr.Width * MULTI, toOcr.Height * MULTI);
Graphics g = Graphics.FromImage(resizer);
g.InterpolationMode = System.Drawing.Drawing2D.InterpolationMode.Bicubic;
g.DrawImage(toOcr, 0, 0, resizer.Width, resizer.Height);
Bitmap bw = Create1bppImage(resizer);
Tesseract.Page p = tesseract.Process(bw);
result = new OcrResults(p.GetText(), bw);
// bmp.Save("test0.jpg");
// resizer.Save("test.jpg");
// bw.Save("testBw.jpg");
bw.Dispose();
resizer.Dispose();
g.Dispose();
p.Dispose();
toOcr.Dispose();
return;
}
public static byte[] PerformOCRTesseract(byte[] image)
{
// Specify that Tesseract use three 3 languages: English, Russian and Vietnamese.
string tesseractLanguages = "rus+eng+vie";
// A path to a folder which contains languages data files and font file "pdf.ttf".
// Language data files can be found here:
// Good and fast: https://github.com/tesseract-ocr/tessdata_fast
// or
// Best and slow: https://github.com/tesseract-ocr/tessdata_best
// Also this folder must have write permissions.
string tesseractData = Path.GetFullPath(@"..\..\tessdata\");
// A path for a temporary PDF file (because Tesseract returns OCR result as PDF document)
string tempFile = Path.Combine(tesseractData, Path.GetRandomFileName());
try
{
using (Tesseract.IResultRenderer renderer = Tesseract.PdfResultRenderer.CreatePdfRenderer(tempFile, tesseractData, true))
{
using (renderer.BeginDocument("Serachablepdf"))
{
using (Tesseract.TesseractEngine engine = new Tesseract.TesseractEngine(tesseractData, tesseractLanguages, Tesseract.EngineMode.Default))
{
engine.DefaultPageSegMode = Tesseract.PageSegMode.Auto;
using (MemoryStream msImg = new MemoryStream(image))
{
System.Drawing.Image imgWithText = System.Drawing.Image.FromStream(msImg);
for (int i = 0; i < imgWithText.GetFrameCount(System.Drawing.Imaging.FrameDimension.Page); i++)
{
imgWithText.SelectActiveFrame(System.Drawing.Imaging.FrameDimension.Page, i);
using (MemoryStream ms = new MemoryStream())
{
imgWithText.Save(ms, System.Drawing.Imaging.ImageFormat.Png);
byte[] imgBytes = ms.ToArray();
using (Tesseract.Pix img = Tesseract.Pix.LoadFromMemory(imgBytes))
{
using (var page = engine.Process(img, "Serachablepdf"))
{
renderer.AddPage(page);
}
}
}
}
}
}
}
}
PdfLoadOptions pl = new PdfLoadOptions();
pl.ShowInvisibleText = true;
pl.PreserveEmbeddedFonts = false;
pl.ConversionMode = PdfConversionMode.Continuous;
DocumentCore dc = DocumentCore.Load(File.OpenRead(tempFile + @".pdf"), pl);
byte[] returnPdf;
using (MemoryStream ms = new MemoryStream())
{
PdfSaveOptions ps = new PdfSaveOptions();
dc.Save(ms, ps);
returnPdf = ms.ToArray();
}
return(returnPdf);
}
catch (Exception e)
{
Console.WriteLine();
Console.WriteLine("Please be sure that you have Language data files (*.traineddata) in your folder \"tessdata\"");
Console.WriteLine("The Language data files can be download from here: https://github.com/tesseract-ocr/tessdata_fast");
Console.ReadKey();
throw new Exception("Error Tesseract: " + e.Message);
}
finally
{
if (File.Exists(tempFile + ".pdf"))
{
File.Delete(tempFile + ".pdf");
}
}
}
static public void Start()
{
capture = new VideoCapture(0);
Window window = new Window("Camera");
using (Mat image = new Mat()) // Frame image buffer
{
// Load the cascade
var haarCascade = new CascadeClassifier(@"haarcascade_russian_plate_number.xml");
Mat mat = new Mat();
Rect[] Rects;
Rect[] Rects2 = new Rect[0];
Mat cut_numberMat = new Mat();
Bitmap cut_numberBitmap;
Bitmap fullPictureBitmap;
Mat fullPictureMat = new Mat();
string numberStr;
char[] numberChar;
Tesseract.TesseractEngine ocr = new Tesseract.TesseractEngine("./tessdata", "eng", Tesseract.EngineMode.Default); //Tesseract.EngineMode.TesseractAndCube
Tesseract.Page page;
// When the movie playback reaches end, Mat.data becomes NULL.
while (true)
{
using (Mat frame = GetFrame(true))
{
if (frame.Empty())
{
break;
}
window.ShowImage(frame);
// Detect number
Rects = haarCascade.DetectMultiScale(frame, 1.1, 4, HaarDetectionType.DoRoughSearch | HaarDetectionType.DoCannyPruning, new OpenCvSharp.Size(100, 100)); //, new OpenCvSharp.Size(100, 100)
if (Rects.Length > 0)
{
MemoryStream stream = new MemoryStream();
Bitmap tmpBitmap = OpenCvSharp.Extensions.BitmapConverter.ToBitmap(frame);
tmpBitmap.Save(stream, System.Drawing.Imaging.ImageFormat.Jpeg);
var pic = Image.FromStream(stream);
using (Graphics g = Graphics.FromImage(pic))
{
g.DrawImage(tmpBitmap, 0, 0);
Pen p = new Pen(Color.Red, 5);
g.DrawRectangle(p, Rects[0].X, Rects[0].Y, Rects[0].Width, Rects[0].Height);
g.Save();
//tmpBitmap.Save("myfile2.png", System.Drawing.Imaging.ImageFormat.Png);
//tmpBitmap.Save(stream, System.Drawing.Imaging.ImageFormat.Jpeg);
fullPictureBitmap = (Bitmap)pic;
Cv2.Canny(frame, fullPictureMat, 150, 300);
mat = OpenCvSharp.Extensions.BitmapConverter.ToMat(fullPictureBitmap);
//Cv2.Canny(frame, fullPictureMat, 50, 200);
fullPictureBitmap = OpenCvSharp.Extensions.BitmapConverter.ToBitmap(fullPictureMat);
Rectangle rectangle = new Rectangle(Rects[0].X, Rects[0].Y, Rects[0].Width, Rects[0].Height);
//Region region = new Region(rectangle);
cut_numberBitmap = fullPictureBitmap.Clone(rectangle, System.Drawing.Imaging.PixelFormat.Format8bppIndexed);
page = ocr.Process(cut_numberBitmap);
numberStr = page.GetText();
if (numberStr.Length > 4 && numberStr.Length < 10)
{
if (Char.IsDigit(numberStr, 1) == true && Char.IsDigit(numberStr, 2) == true && Char.IsDigit(numberStr, 3) == true)
{
cut_numberBitmap.Save("myfile2.png", System.Drawing.Imaging.ImageFormat.Png);
Console.WriteLine(numberStr);
}
}
page.Dispose();
//cut_numberBitmap.Dispose();
}
window.ShowImage(mat);
}
}
Cv2.WaitKey(30);
}
}
}