public void CanDetectOrientationForMode(
[Values(PageSegMode.Auto,
PageSegMode.AutoOnly,
PageSegMode.AutoOsd,
PageSegMode.CircleWord,
PageSegMode.OsdOnly,
PageSegMode.SingleBlock,
PageSegMode.SingleBlockVertText,
PageSegMode.SingleChar,
PageSegMode.SingleColumn,
PageSegMode.SingleLine,
PageSegMode.SingleWord)]
PageSegMode pageSegMode)
{
using (var img = LoadTestImage(ExampleImagePath)) {
using (var rotatedPix = img.Rotate((float)Math.PI)) {
using (var page = engine.Process(rotatedPix, pageSegMode)) {
int orientation;
float confidence;
string scriptName;
float scriptConfidence;
page.DetectBestOrientationAndScript(out orientation, out confidence, out scriptName, out scriptConfidence);
Assert.That(orientation, Is.EqualTo(180));
Assert.That(scriptName, Is.EqualTo("Latin"));
}
}
}
}
public List <CharBox> GetCharBoxs(Bitmap b, PageSegMode pageSegMode)
{
using (Tesseract.Page page = engine.Process(b, pageSegMode))
{
return(getCharBoxs(page));
}
}
public string GetHtml(Bitmap b, PageSegMode pageSegMode)
{
using (Tesseract.Page page = engine.Process(b, pageSegMode))
{
return(page.GetHOCRText(0, false));
}
}
/// <summary> Use Tesseract OCR to find words on picture to string </summary>
public static string AnalyzeText(Bitmap bitmap, PageSegMode pageMode = PageSegMode.SingleLine, bool numbersOnly = false)
{
string text = "";
TesseractEngine e;
while (!engines.TryTake(out e))
{
Thread.Sleep(10);
}
if (numbersOnly)
{
e.SetVariable("tessedit_char_whitelist", "0123456789");
}
using (var page = e.Process(bitmap, pageMode))
{
using (var iter = page.GetIterator())
{
iter.Begin();
do
{
text += iter.GetText(PageIteratorLevel.TextLine);
}while (iter.Next(PageIteratorLevel.TextLine));
}
}
engines.Add(e);
return(text);
}
public void CanDetectOrientationForMode(
[Values(PageSegMode.Auto,
PageSegMode.AutoOnly,
PageSegMode.AutoOsd,
PageSegMode.CircleWord,
PageSegMode.OsdOnly,
PageSegMode.SingleBlock,
PageSegMode.SingleBlockVertText,
PageSegMode.SingleChar,
PageSegMode.SingleColumn,
PageSegMode.SingleLine,
PageSegMode.SingleWord)]
PageSegMode pageSegMode)
{
using (var img = Pix.LoadFromFile(@".\phototest.tif")) {
using (var rotatedPix = img.Rotate((float)Math.PI)) {
using (var page = engine.Process(rotatedPix, pageSegMode)) {
Orientation orientation;
float confidence;
page.DetectBestOrientation(out orientation, out confidence);
Assert.That(orientation, Is.EqualTo(Orientation.PageDown));
}
}
}
}
internal Page(TesseractEngine engine, Pix image, string imageName, Rect regionOfInterest, PageSegMode pageSegmentMode)
{
Engine = engine;
Image = image;
ImageName = imageName;
RegionOfInterest = regionOfInterest;
PageSegmentMode = pageSegmentMode;
}
internal Page(TesseractEngine engine, Pix image, string imageName, Rect regionOfInterest, PageSegMode pageSegmentMode)
{
Engine = engine;
Image = image;
ImageName = imageName;
RegionOfInterest = regionOfInterest;
PageSegmentMode = pageSegmentMode;
}
public object Convert(object value, Type targetType, object parameter, CultureInfo culture)
{
if (value is PageSegMode)
{
PageSegMode mode = (PageSegMode)value;
return($"{(int)mode}.{mode.ToString()}");
}
return(null);
}
public void SetPageSegMode(PageSegMode mode)
{
if (this.tesseract == null)
{
this.tesseract = new Tesseract();
}
tesseract.set_value(mode);
}
//получение номера серии через с помощью масштабирования страницы с разным шагом(исп-ть, если не находит на исходном масштабе)
//sourceImage = входное изображение, stepScale = шаг масштабирования, amountOfSteps = кол-во проходов рескейлинга
public string getSeriaNumberViaRescaling(Image sourceImage, float stepScale, int amountOfSteps)
{
//подключаем распознаватель
TesseractEngine ocrRus = new TesseractEngine("./tessdata", "rus", EngineMode.Default);
TesseractEngine ocrEng = new TesseractEngine("./tessdata", "eng", EngineMode.Default);
//создаем графический фильтр
TResizeTool resizeTool = new TResizeTool();
//ставим методику распознавания страницы
PageSegMode pSegMode = PageSegMode.SparseText;
string seriaNumber = "";
float scaleValue = 1.0f;//исходный масштаб
List <string> seriaList = new List <string>();
for (int k = 1; k <= amountOfSteps; k++)
{
//по индексу отмасш-й страницы берем исходную(без масштаба)
Image img = sourceImage;
int newWidth = Convert.ToInt32((img.Width) * ((k * stepScale) + scaleValue));
int newHeight = Convert.ToInt32((img.Height) * ((k * stepScale) + scaleValue));
Bitmap imgBitmap = resizeTool.resizeImage(img, newWidth, newHeight);
Image rescaledPage = ((Image)imgBitmap);
string rescaledPageTextRus = "";
string rescaledPageTextEng = "";
//удаляем процесс, чтобы не было ошибки
using (var pageRus = ocrRus.Process(imgBitmap, pSegMode))
{
rescaledPageTextRus = pageRus.GetText();
}
//удаляем процесс, чтобы не было ошибки
using (var pageEng = ocrEng.Process(imgBitmap, pSegMode))
{
rescaledPageTextEng = pageEng.GetText();
}
seriaList.Add(getSeriaNumber(rescaledPageTextRus, rescaledPageTextEng));
}
seriaNumber = highestMatch(seriaList);
ocrRus.Dispose();
ocrEng.Dispose();
return(seriaNumber);
}
public string GetTextSurroundedByRectangle(Bitmap b, RectangleF r, PageSegMode pageSegMode)
{
if (!getScaled(b, ref r))
{
return(null);
}
using (Tesseract.Page page = engine.Process(b, new Rect((int)r.X, (int)r.Y, (int)r.Width, (int)r.Height), pageSegMode))
{
return(page.GetText());
}
}
public void CanParseText_UsingMode(PageSegMode mode, String expectedText)
{
using (var engine = CreateEngine()) {
var demoFilename = String.Format("./Ocr/PSM_{0}.png", mode);
using (var pix = LoadTestPix(demoFilename)) {
using (var page = engine.Process(pix, mode)) {
var text = page.GetText().Trim();
Assert.That(text, Is.EqualTo(expectedText));
}
}
}
}
public void CanParseText_UsingMode(PageSegMode mode, String expectedText)
{
using (var engine = new TesseractEngine(@"./tessdata", "eng", EngineMode.Default)) {
var demoFilename = String.Format("./Data/Ocr/PSM_{0}.png", mode);
using (var pix = Pix.LoadFromFile(demoFilename)) {
using (var page = engine.Process(pix, mode)) {
var text = page.GetText().Trim();
Assert.That(text, Is.EqualTo(expectedText));
}
}
}
}
public void CanParseText_UsingMode(PageSegMode mode, String expectedText)
{
using (var engine = new TesseractEngine(@"./tessdata", "eng", EngineMode.Default)) {
var demoFilename = String.Format("./Data/Ocr/PSM_{0}.png", mode);
using (var pix = Pix.LoadFromFile(demoFilename)) {
using (var page = engine.Process(pix, mode)) {
var text = page.GetText().Trim();
Assert.That(text, Is.EqualTo(expectedText));
}
}
}
}
//[DataRow(PageSegMode.SingleBlockVertText, "A line of text", Ignore = "Vertical data missing")]
public void CanParseText_UsingMode(PageSegMode mode, string expectedText)
{
using (var engine = CreateEngine(mode: EngineMode.TesseractAndLstm))
{
var demoFilename = string.Format("./Ocr/PSM_{0}.png", mode);
using (var pix = LoadTestPix(demoFilename))
{
using (var page = engine.Process(pix, mode))
{
var text = page.GetText().Trim();
Assert.AreEqual(text, expectedText);
}
}
}
}
/// <summary>
/// Processes the specific image.
/// </summary>
/// <remarks>
/// You can only have one result iterator open at any one time.
/// </remarks>
/// <param name="image">The image to process.</param>
/// <param name="region">The image region to process.</param>
/// <returns>A result iterator</returns>
public Page Process(Pix image, Rect region, PageSegMode? pageSegMode = null)
{
if (image == null) throw new ArgumentNullException("image");
if (region.X1 < 0 || region.Y1 < 0 || region.X2 > image.Width || region.Y2 > image.Height)
throw new ArgumentException("The image region to be processed must be within the image bounds.", "region");
if (processCount > 0) throw new InvalidOperationException("Only one image can be processed at once. Please make sure you dispose of the page once your finished with it.");
processCount++;
Interop.TessApi.BaseAPISetPageSegMode(handle, pageSegMode.HasValue ? pageSegMode.Value : DefaultPageSegMode);
Interop.TessApi.BaseApiSetImage(handle, image.Handle);
Interop.TessApi.BaseApiSetRectangle(handle, region.X1, region.Y1, region.Width, region.Height);
var page = new Page(this);
page.Disposed += OnIteratorDisposed;
return page;
}
/// <summary>
/// 识别白底黑字的二值化字符图片
/// </summary>
/// <param name="bmp">白底黑字的二值化字符图片</param>
/// <param name="mode">页面分析模式</param>
/// <returns>识别出的字符串</returns>
protected virtual string CrackCodeCore(Bitmap bmp, PageSegMode mode)
{
if (_engine == null)
{
throw new InvalidOperationException("Tesseract识别引擎 未初始化");
}
Tesseract.PageSegMode tmode = mode.CastTo(Tesseract.PageSegMode.Auto);
using (Page page = _engine.Process(bmp, tmode))
{
string text = page.GetText().Replace("\n", "").Replace(" ", "");
if (mode == PageSegMode.SingleBlockVertText)
{
text = text.ReverseString();
}
return(text);
}
}
private string tesseractGetOCR(string fileName, string tessDir, PageSegMode pageMode = PageSegMode.SingleLine)
{
//Console.WriteLine($"tessdir {tessDir}");
using (var engine = new TesseractEngine(tessDir, "eng", EngineMode.Default, "bank.citic"))
{
using (var img = Pix.LoadFromFile(fileName))
{
using (var page = engine.Process(img, pageMode))
{
var text = page.GetText();
text = text.Replace(" ", "");
text = text.Replace("\t", "");
text = text.Replace("\n", "");
text = text.Replace("\r", "");
text = text.Replace("~", "");
return(text.Trim());
}
}
}
}
/// <summary>
/// ocr 识别验证码
/// </summary>
/// <param name="fileName"></param>
/// <param name="pageMode"></param>
/// <returns></returns>
private string OCRGetStringByImage(string fileName, PageSegMode pageMode = PageSegMode.SingleLine)
{
string dir = System.Windows.Forms.Application.StartupPath;
string tessDir = $"{dir}\\tessdata";
using (var engine = new TesseractEngine(tessDir, "eng", EngineMode.Default, "bank.abc"))
{
using (var img = Pix.LoadFromFile(fileName))
{
using (var page = engine.Process(img, pageMode))
{
var text = page.GetText();
text = text.Replace(" ", "");
text = text.Replace("\t", "");
text = text.Replace("\n", "");
text = text.Replace("\r", "");
text = text.Replace("~", "");
return(text.Trim());
}
}
}
}
/// <summary>
/// Recognize the license plate number from a given image
/// </summary>
/// <remarks>
/// Uses a Tesseract OCR library to recognize the set of characters.
/// When you would like to distinguish single character PageSegMode.SingleChar should be used,
/// otherwise PageSegMode.SingleBlock will be applied.
/// Tesseract uses already prepared training data which consists of built-in data set and
/// special training set for polish license plates.
/// </remarks>
/// <param name="imgWithNumber">Mat containing the image of possible license plate area</param>
/// <param name="pageMode">PageSegMode which should be used when recognizing the character </param>
/// <returns>Recognized plate number</returns>
private string RecognizeNumber(Mat imgWithNumber, PageSegMode pageMode = PageSegMode.SingleChar)
{
Tesseract.Character[] characters;
StringBuilder licensePlateNumber = new StringBuilder();
using (var ocr = new Tesseract())
{
ocr.Init(
_ocrParams["TEST_DATA_PATH"],
_ocrParams["TEST_DATA_LANG"],
OcrEngineMode.LstmOnly);
ocr.SetVariable(
"tessedit_char_whitelist",
_ocrParams["WHITE_LIST"]);
ocr.SetVariable
("user_defined_dpi", "70");
ocr.PageSegMode = pageMode;
using (Mat tmp = imgWithNumber.Clone())
{
ocr.SetImage(tmp);
ocr.Recognize();
characters = ocr.GetCharacters();
for (int i = 0; i < characters.Length; i++)
{
licensePlateNumber.Append(characters[i].Text);
}
}
return(licensePlateNumber.ToString());
}
}
public string OCRProcess(Bitmap originImg)
{
try {
Bitmap cleanImg = originImg;
if (!IsBlackFontWhiteBack(originImg))
{
cleanImg = GetCleanImg(originImg);
}
PageSegMode pageSegMode = PageSegMode.Auto;
if (originImg.Width < 100)
{
pageSegMode = PageSegMode.SingleWord;
}
using (var page = TessOCR.Process(cleanImg, pageSegMode))
{
string res = page.GetText();
res = res.Trim();
if (res.Equals(String.Empty))
{
saveFailImg(originImg);
saveFailImg(cleanImg);
}
page.Dispose();
return(res);
}
}
catch (Exception ex) {
errorInfo = ex.Message;
File.AppendAllText("log.txt", String.Format("{0} {1}", DateTime.Now.ToString(), ex.Message));
}
return(String.Empty);
}
/// <summary>
/// Process the specified bitmap image.
/// </summary>
/// <remarks>
/// Please consider <see cref="Process(Pix, PageSegMode?)"/> instead. This is because
/// this method must convert the bitmap to a pix for processing which will add additional overhead.
/// Leptonica also supports a large number of image pre-processing functions as well.
/// </remarks>
/// <param name="image">The image to process.</param>
/// <param name="pageSegMode">The page segmentation mode.</param>
/// <returns></returns>
public Page Process(Bitmap image, PageSegMode? pageSegMode = null)
{
return Process(image, new Rect(0, 0, image.Width, image.Height), pageSegMode);
}
public static extern void BaseAPISetPageSegMode(IntPtr handle, PageSegMode mode);
public void SetPageSegMode(PageSegMode mode)
{
if (this.tesseract == null)
{
this.tesseract = new Tesseract();
}
tesseract.set_value(mode);
}
private void ocrImage(ref List <Image> imagePdfPages, ref List <Image> imageScaledPages, ref List <pdfPageInfo> infoPages, object indexOfPage)
{
///////////////////////////////////////////////////////////////////////////
//распознаем изображение
TesseractEngine ocrRus = new TesseractEngine("./tessdata", "rus", EngineMode.Default);
TesseractEngine ocrEng = new TesseractEngine("./tessdata", "eng", EngineMode.Default);
Regex regexFullNumberRus = new Regex(@"ТС[^\\n]+", RegexOptions.IgnoreCase | RegexOptions.Singleline);
Regex regexFullNumberEng = new Regex(@"TC[^\\n]+", RegexOptions.IgnoreCase | RegexOptions.Singleline);
//Regex regexOrgSertTypeSeriaShortNumberRus = new Regex(@"[А-Я]{2}[0-9]{2}\W[0-9]{5}", RegexOptions.IgnoreCase | RegexOptions.Singleline);
Regex regexOrgSertTypeSeriaShortNumberRus = new Regex(@"[А-Я]{2}[0-9]{2}[А-Я]{1}[0-9]{5}", RegexOptions.IgnoreCase | RegexOptions.Singleline);
//Regex regexSeriaNumber = new Regex(@"[0-9]{7}", RegexOptions.IgnoreCase | RegexOptions.Singleline | RegexOptions.RightToLeft);
Regex regexSeriaNumber = new Regex(@"[0-9]{7}", RegexOptions.IgnoreCase);
//берем страницу
Bitmap imgBitmapOcrRus = (Bitmap)imageScaledPages[(int)indexOfPage];
Bitmap imgBitmapOcrEng = (Bitmap)imageScaledPages[(int)indexOfPage];
//ставим методику распознавания страницы
PageSegMode pSegMode = PageSegMode.SparseText;
string pdfTextRus = "";
string pdfTextEng = "";
//waitHandler.WaitOne();
//удаляем процесс, чтобы не было ошибки
using (var pageRus = ocrRus.Process(imgBitmapOcrRus, pSegMode)) {
pdfTextRus = pageRus.GetText();
}
//удаляем процесс, чтобы не было ошибки
using (var pageEng = ocrEng.Process(imgBitmapOcrEng, pSegMode)) {
pdfTextEng = pageEng.GetText();
}
//waitHandler.Set();
if (isSertificate(pdfTextRus))//сертификат
//if (abbyOCR.isSertificateViaRescaling(imagePdfPages[(int)indexOfPage], 0.1f, 3))//сертификат
{
//////////////
//tesOCR.getDataViaRescaling(imagePdfPages[j], ref infoPage, ref words, 0.1f, 10);
pdfPageInfo infoPageABBY = abbyOCR.getDataViaRescaling(ref ocrTasks[(int)indexOfPage], imagePdfPages[(int)indexOfPage], 0.1f, 4);
infoPageABBY.numberOfPDFPage_ = (int)indexOfPage;
//////////////////////////////////////////////////
//infoPages[(int)indexOfPage]=infoPageABBY;
if ((int)indexOfPage != 0)
{
ocrTasks[(int)indexOfPage - 1].Wait();
}
infoPages.Insert((int)indexOfPage, infoPageABBY);
}
else //приложение
//ждем завершения предыдущего такска, чтобы взять оттуда инфу
//ocrTasks[(int)indexOfPage - 1].Wait();
{
pdfPageInfo infoPageABBY = abbyOCR.getDataViaRescaling(ref ocrTasks[(int)indexOfPage], imagePdfPages[(int)indexOfPage], 0.1f, 4);
infoPageABBY.numberOfPDFPage_ = (int)indexOfPage;
infoPageABBY.typeOfPage_ = typeOfPage.PRILOZENIE;
//////////////////////////////////////////////////
//infoPages[(int)indexOfPage]=infoPageABBY;
if ((int)indexOfPage != 0)
{
ocrTasks[(int)indexOfPage - 1].Wait();
}
infoPages.Insert((int)indexOfPage, infoPageABBY);
/*//ждем завершения предыдущего такска, чтобы взять оттуда инфу
* ocrTasks[(int)indexOfPage - 1].Wait();
*
* pdfPageInfo infoPage = infoPages[infoPages.Count - 1];
*
* //если серт не распознали, то и приложение не распознаем
* if (infoPage.isTesseracted_ == false) {
* infoPage.numberOfPDFPage_ = (int)indexOfPage;
* infoPage.isTesseracted_ = false;
* infoPage.typeOfPage_ = typeOfPage.PRILOZENIE;
* infoPage.textRus_ = pdfTextRus;
* infoPage.textEng_ = pdfTextEng;
* infoPage.seriaNumber_ = "";
*
* //infoPage.ocrMethod_ = ocrMethod.TESSERACT;
* infoPage.ocrMethod_ = ocrMethod.ABBY;
*
* infoPage.fullNumber_ = infoPage.techReglament_ + infoPage.typeOfSert_ + "-" + infoPage.country_ + "." + infoPage.orgSert_ + "." + infoPage.typeSeria_ + "." + infoPage.shortNumber_ + "." + infoPage.seriaNumber_;
* infoPage.fullNumber_ = infoPage.fullNumber_.ToUpper();//ставим верхний регистр
*
* //infoPages[(int)indexOfPage] = infoPage;
* infoPages.Insert((int)indexOfPage, infoPage);
*
* //освобождаем ресурсы
* imgBitmapOcrRus.Dispose();
* imgBitmapOcrEng.Dispose();
* }
*
* //если до этого было приложение, то seriaNumber+1
* if (infoPage.typeOfPage_ == typeOfPage.PRILOZENIE) {
* infoPage.numberOfPDFPage_ = (int)indexOfPage;
* infoPage.textRus_ = pdfTextRus;
* infoPage.textEng_ = pdfTextEng;
*
* int sertSeriaNumber = Convert.ToInt32(infoPage.seriaNumber_);
* string clearSeriaNumber = "0000000";
* string prilSeriaNumber = Convert.ToString(sertSeriaNumber + 1);
*
* int amountOfSymbolsSSN = prilSeriaNumber.Length;
* int amountClearSN = clearSeriaNumber.Length;
*
* //clearSeriaNumber=clearSeriaNumber.Remove(2, 5).Insert(2, prilSeriaNumber);
* clearSeriaNumber = clearSeriaNumber.Remove(amountClearSN - amountOfSymbolsSSN, amountOfSymbolsSSN).Insert(amountClearSN - amountOfSymbolsSSN, prilSeriaNumber);
*
* infoPage.seriaNumber_ = clearSeriaNumber;
* infoPage.fullNumber_ = infoPage.techReglament_ + infoPage.typeOfSert_ + "-" + infoPage.country_ + "." + infoPage.orgSert_ + "." + infoPage.typeSeria_ + "." + infoPage.shortNumber_ + "." + infoPage.seriaNumber_;
* //infoPage.ocrMethod_ = ocrMethod.TESSERACT;
* infoPage.ocrMethod_ = ocrMethod.ABBY;
* infoPage.isTesseracted_ = true;
*
* infoPage.fullNumber_ = infoPage.fullNumber_.ToUpper();//ставим верхний регистр
*
* //infoPages[(int)indexOfPage] = infoPage;
* infoPages.Insert((int)indexOfPage, infoPage);
* }
*
* infoPage.numberOfPDFPage_ = (int)indexOfPage;
* infoPage.isTesseracted_ = false;
* infoPage.typeOfPage_ = typeOfPage.PRILOZENIE;
* //infoPage.ocrMethod_ = ocrMethod.TESSERACT;
* infoPage.ocrMethod_ = ocrMethod.ABBY;
* infoPage.textRus_ = pdfTextRus;
* infoPage.textEng_ = pdfTextEng;
*
* //infoPage.seriaNumber_ = tesOCR.getSeriaNumber(pdfTextRus, pdfTextEng);
* infoPage.seriaNumber_ = abbyOCR.getSeriaNumberViaRescaling(imagePdfPages[(int)indexOfPage], 0.1f, 3);
* infoPage.fullNumber_ = infoPage.techReglament_ + infoPage.typeOfSert_ + "-" + infoPage.country_ + "." + infoPage.orgSert_ + "." + infoPage.typeSeria_ + "." + infoPage.shortNumber_ + "." + infoPage.seriaNumber_;
*
* if (infoPage.seriaNumber_ == "") {
* infoPage.isTesseracted_ = false;
* } else {
* infoPage.isTesseracted_ = true;
* }
*
* infoPage.fullNumber_ = infoPage.fullNumber_.ToUpper();//ставим верхний регистр
*
* //infoPages[(int)indexOfPage] = infoPage;
* infoPages.Insert((int)indexOfPage, infoPage);*/
}
//освобождаем ресурсы
imgBitmapOcrRus.Dispose();
imgBitmapOcrEng.Dispose();
}
/// <inheritdoc />
public void BaseApiSetPageSegMode(IEngineHandle engine,
PageSegMode mode
)
{
UnsafeTesseractPrimitives.BaseAPISetPageSegMode(engine.Handle, (int)mode);
}
public static extern void BaseAPISetPageSegMode(IntPtr handle, PageSegMode mode);
/// <summary>
/// Processes the specific image.
/// </summary>
/// <remarks>
/// You can only have one result iterator open at any one time.
/// </remarks>
/// <param name="image">The image to process.</param>
/// <param name="inputName">Sets the input file's name, only needed for training or loading a uzn file.</param>
/// <param name="pageSegMode">The page layout analyasis method to use.</param>
public Page Process(Pix image, string inputName, PageSegMode? pageSegMode = null)
{
return Process(image, inputName, new Rect(0, 0, image.Width, image.Height), pageSegMode);
}
public void getDataViaRescaling(Image sourceImage, ref Tesseract_OCR_Window.pdfPageInfo infoPage, ref Dictionary <string, string> replaceWords, float stepScale, int amountOfSteps)
{
//подключаем распознаватель
TesseractEngine ocrRus = new TesseractEngine("./tessdata", "rus", EngineMode.Default);
TesseractEngine ocrEng = new TesseractEngine("./tessdata", "eng", EngineMode.Default);
//создаем графический фильтр
TResizeTool resizeTool = new TResizeTool();
//ставим методику распознавания страницы
PageSegMode pSegMode = PageSegMode.SparseText;
float scaleValue = 1.0f;//исходный масштаб
List <string> seriaNumberList = new List <string>();
List <string> countryList = new List <string>();
List <string> orgSertList = new List <string>();
List <string> shortNumberList = new List <string>();
List <string> typeSeriaList = new List <string>();
for (int k = 0; k <= amountOfSteps; k++)
{
//по индексу отмасш-й страницы берем исходную(без масштаба)
Image img = sourceImage;
int newWidth = Convert.ToInt32((img.Width) * ((k * stepScale) + scaleValue));
int newHeight = Convert.ToInt32((img.Height) * ((k * stepScale) + scaleValue));
Bitmap imgBitmap = resizeTool.resizeImage(img, newWidth, newHeight);
Image rescaledPage = ((Image)imgBitmap);
string rescaledPageTextRus = "";
string rescaledPageTextEng = "";
//удаляем процесс, чтобы не было ошибки
using (var pageRus = ocrRus.Process(imgBitmap, pSegMode))
{
rescaledPageTextRus = pageRus.GetText();
}
//удаляем процесс, чтобы не было ошибки
using (var pageEng = ocrEng.Process(imgBitmap, pSegMode))
{
rescaledPageTextEng = pageEng.GetText();
}
//набираем статистику для последующего поиска наибольшего соответствия
string seriaNumber = getSeriaNumber(rescaledPageTextRus, rescaledPageTextEng);
string country = getCountry(rescaledPageTextEng);
string orgSert = getOrgSert(rescaledPageTextRus);
string shortNumber = getShortNumber(rescaledPageTextRus);
string typeSeria = getTypeSeria(rescaledPageTextRus, rescaledPageTextEng, ref replaceWords);
if (seriaNumber != "")
{
seriaNumberList.Add(seriaNumber);//серия
}
if (country != "")
{
countryList.Add(country);//страна
}
if (orgSert != "")
{
orgSertList.Add(orgSert);//орган по сертификации
}
if (shortNumber != "")
{
shortNumberList.Add(shortNumber);//shortNumber
}
if (typeSeria != "")
{
typeSeriaList.Add(typeSeria);//тип серии
}
}
ocrRus.Dispose();
ocrEng.Dispose();
//собираем данные для структуры infoPage
infoPage.seriaNumber_ = highestMatch(seriaNumberList);
infoPage.country_ = highestMatch(countryList);
infoPage.orgSert_ = highestMatch(orgSertList);
infoPage.shortNumber_ = highestMatch(shortNumberList);
infoPage.typeSeria_ = highestMatch(typeSeriaList);
}
public void set_value(PageSegMode mode)
{
this.mode = mode;
}
internal static extern void BaseAPISetPageSegMode(HandleRef handle, PageSegMode mode);
public static extern void BaseAPISetPageSegMode(HandleRef handle, PageSegMode mode);
/// <summary>
/// 识别验证码
/// </summary>
/// <param name="bmp">要识别的图像</param>
/// <param name="mode">页面分析模式</param>
/// <returns>识别出的字符串</returns>
public string CrackCode(Bitmap bmp, PageSegMode mode)
{
Bitmap binBmp = Binary(bmp);
return(CrackCodeCore(binBmp, mode));
}
static void Main(string[] args)
{
if (args.Length == 0)
{
printUsage("");
return;
}
if ((args.Length == 1 && args[0] == "-v") || (args.Length == 1 && args[0] == "--version"))
{
Console.WriteLine(string.Format("tesseract-{0}", TessBaseAPI.Version()));
Console.ReadKey();
return;
}
string lang = "eng";
string image = "";
string output = "";
PageSegMode pagesegmode = PageSegMode.PSM_AUTO;
int arg = 0;
int argc = args.Length;
while (arg < argc && (output == string.Empty || args[arg].StartsWith("-")))
{
if (args[arg] == "-l" && arg + 1 < argc)
{
lang = args[arg + 1];
++arg;
}
else if (args[arg] == "-psm" && arg + 1 < argc)
{
//assign pagesegmode
++arg;
}
else if (image == string.Empty)
{
image = args[arg];
}
else if (output == string.Empty)
{
output = args[arg];
}
++arg;
}
if (string.IsNullOrEmpty(output))
{
printUsage(args[0]);
return;
}
TessBaseAPI api = new TessBaseAPI();
api.SetOutputName(output);
api.SetPageSegMode(pagesegmode);
api.Init(args[0],
lang,
OcrEngineMode.OEM_DEFAULT,
args[arg - 1],
argc - arg,
null,
null,
false);
Console.WriteLine(string.Format("Tesseract Open Source OCR Engine v{0} with Leptonica\n",
TessBaseAPI.Version()));
string text_out = string.Empty;
if (!api.ProcessPages(image, null, 0, ref text_out))
{
Console.WriteLine("Error during processing.\n");
}
bool output_hocr = false;
api.GetBoolVariable("tessedit_create_hocr", ref output_hocr);
bool output_box = false;
api.GetBoolVariable("tessedit_create_boxfile", ref output_box);
string outfile = output;
outfile += output_hocr ? ".html" : output_box ? ".box" : ".txt";
File.WriteAllText(outfile, text_out);
}
/// <summary>
/// Process the specified bitmap image.
/// </summary>
/// <remarks>
/// Please consider <see cref="TesseractEngine.Process(Pix, String, Rect, PageSegMode?)"/> instead. This is because
/// this method must convert the bitmap to a pix for processing which will add additional overhead.
/// Leptonica also supports a large number of image pre-processing functions as well.
/// </remarks>
/// <param name="image">The image to process.</param>
/// <param name="inputName">Sets the input file's name, only needed for training or loading a uzn file.</param>
/// <param name="region">The region of the image to process.</param>
/// <param name="pageSegMode">The page segmentation mode.</param>
/// <returns></returns>
public Page Process(Bitmap image, string inputName, Rect region, PageSegMode? pageSegMode = null)
{
using (var pix = PixConverter.ToPix(image)) {
return Process(pix, inputName, region, pageSegMode);
}
}
/// <summary>
/// Processes a specified region in the image using the specified page layout analysis mode.
/// </summary>
/// <remarks>
/// You can only have one result iterator open at any one time.
/// </remarks>
/// <param name="image">The image to process.</param>
/// <param name="region">The image region to process.</param>
/// <param name="pageSegMode">The page layout analyasis method to use.</param>
/// <returns>A result iterator</returns>
public Page Process(Pix image, Rect region, PageSegMode? pageSegMode = null)
{
return Process(image, null, region, pageSegMode);
}
internal static extern void TessBaseAPISetPageSegMode(IntPtr ocr, PageSegMode mode);
/// <summary>
/// 图片文字识别
/// </summary>
/// <param name="imageUrl">图片路径</param>
/// <param name="tessdataUrl">语言包路径</param>
/// <param name="language">语言包名称</param>
/// <param name="variableData">设置识别变量，默认识别数字</param>
/// <param name="pageSegMode">PageSegMode</param>
/// <returns></returns>
public static string Identity(string imageUrl, string tessdataUrl, string language = "eng", string variableData = "0123456789", PageSegMode pageSegMode = PageSegMode.Auto)
{
Bitmap bitmap = null;
TesseractEngine tesseractEngine = null;
Page page = null;
try
{
bitmap = new Bitmap(imageUrl);
tesseractEngine = new TesseractEngine(tessdataUrl, language, EngineMode.Default);
if (!string.IsNullOrEmpty(variableData))
{
tesseractEngine.SetVariable("tessedit_char_whitelist", variableData);
}
page = tesseractEngine.Process(PixConverter.ToPix(bitmap), pageSegMode);
return(page.GetText());
}
catch
{
throw;
}
finally
{
if (bitmap != null)
{
bitmap.Dispose();
}
if (page != null)
{
page.Dispose();
}
if (tesseractEngine != null)
{
tesseractEngine.Dispose();
}
}
}
/// <summary>
/// Processes a specified region in the image using the specified page layout analysis mode.
/// </summary>
/// <remarks>
/// You can only have one result iterator open at any one time.
/// </remarks>
/// <param name="image">The image to process.</param>
/// <param name="inputName">Sets the input file's name, only needed for training or loading a uzn file.</param>
/// <param name="region">The image region to process.</param>
/// <param name="pageSegMode">The page layout analyasis method to use.</param>
/// <returns>A result iterator</returns>
public Page Process(Pix image, string inputName, Rect region, PageSegMode? pageSegMode = null)
{
if (image == null) throw new ArgumentNullException("image");
if (region.X1 < 0 || region.Y1 < 0 || region.X2 > image.Width || region.Y2 > image.Height)
throw new ArgumentException("The image region to be processed must be within the image bounds.", "region");
if (processCount > 0) throw new InvalidOperationException("Only one image can be processed at once. Please make sure you dispose of the page once your finished with it.");
processCount++;
var actualPageSegmentMode = pageSegMode.HasValue ? pageSegMode.Value : DefaultPageSegMode;
Interop.TessApi.Native.BaseAPISetPageSegMode(handle, actualPageSegmentMode);
Interop.TessApi.Native.BaseApiSetImage(handle, image.Handle);
if(!String.IsNullOrEmpty(inputName)) {
Interop.TessApi.Native.BaseApiSetInputName(handle, inputName);
}
var page = new Page(this, image, inputName, region, actualPageSegmentMode);
page.Disposed += OnIteratorDisposed;
return page;
}
//!!!abandoned for GetCharBoxsSurroundedByRectangle() because:
//- it does not give good end lines;
//- it does not accept expilicitly TextAutoInsertSpace
//public string GetTextSurroundedByRectangle(Bitmap b, RectangleF r, PageSegMode pageSegMode)
//{
// if (!getScaled(b, ref r))
// return string.Empty;
// using (Tesseract.Page page = engine.Process(b, new Rect((int)r.X, (int)r.Y, (int)r.Width, (int)r.Height), pageSegMode))
// {
// return page.GetText();
// }
//}
public List <CharBox> GetCharBoxsSurroundedByRectangle(Bitmap b, RectangleF r, PageSegMode pageSegMode)
{
if (!getScaled(b, ref r))
{
return(null);
}
using (Tesseract.Page page = engine.Process(b, new Rect((int)r.X, (int)r.Y, (int)r.Width, (int)r.Height), pageSegMode))
{
return(getCharBoxs(page));
}
}
//получение всех данные по странице с помощью масштабирования страницы с разным шагом(исп-ть, если не находит на исходном масштабе)
//sourceImage = входное изображение, stepScale = шаг масштабирования, amountOfSteps = кол-во проходов рескейлинга,
//infoPage = ссылка на структуру для заполнения
public void getData(Image sourceImage, ref Tesseract_OCR_Window.pdfPageInfo infoPage, ref Dictionary <string, string> replaceWords, float stepScale, int amountOfSteps)
{
//подключаем распознаватель
TesseractEngine ocrRus = new TesseractEngine("./tessdata", "rus", EngineMode.Default);
TesseractEngine ocrEng = new TesseractEngine("./tessdata", "eng", EngineMode.Default);
//создаем графический фильтр
TResizeTool resizeTool = new TResizeTool();
//ставим методику распознавания страницы
PageSegMode pSegMode = PageSegMode.SparseText;
string seriaNumber = ""; //серия
string country = ""; //страна
string orgSert = ""; //организация
string shortNumber = ""; //short number
string typeSeria = ""; //тип серии
float scaleValue = 1.0f; //исходный масштаб
for (int k = 0; k <= amountOfSteps; k++)
{
//по индексу отмасш-й страницы берем исходную(без масштаба)
Image img = sourceImage;
int newWidth = Convert.ToInt32((img.Width) * ((k * stepScale) + scaleValue));
int newHeight = Convert.ToInt32((img.Height) * ((k * stepScale) + scaleValue));
Bitmap imgBitmap = resizeTool.resizeImage(img, newWidth, newHeight);
Image rescaledPage = ((Image)imgBitmap);
string rescaledPageTextRus = "";
string rescaledPageTextEng = "";
//удаляем процесс, чтобы не было ошибки
using (var pageRus = ocrRus.Process(imgBitmap, pSegMode))
{
rescaledPageTextRus = pageRus.GetText();
}
//удаляем процесс, чтобы не было ошибки
using (var pageEng = ocrEng.Process(imgBitmap, pSegMode))
{
rescaledPageTextEng = pageEng.GetText();
}
//собираем данные для структуры infoPage
if (infoPage.seriaNumber_ == "")
{
seriaNumber = getSeriaNumber(rescaledPageTextRus, rescaledPageTextEng);//серия
if (seriaNumber != "")
{
infoPage.seriaNumber_ = seriaNumber;
}
}
if (infoPage.country_ == "")
{
country = getCountry(rescaledPageTextEng);
if (country != "")
{
infoPage.country_ = country;
}
}
if (infoPage.orgSert_ == "")
{
orgSert = getOrgSert(rescaledPageTextRus);
if (orgSert != "")
{
infoPage.orgSert_ = orgSert;
}
}
if (infoPage.shortNumber_ == "")
{
shortNumber = getShortNumber(rescaledPageTextRus);
if (shortNumber != "")
{
infoPage.shortNumber_ = shortNumber;
}
}
if (infoPage.typeSeria_ == "")
{
typeSeria = getTypeSeria(rescaledPageTextRus, rescaledPageTextEng, ref replaceWords);
if (typeSeria != "")
{
infoPage.typeSeria_ = typeSeria;
}
}
}
ocrRus.Dispose();
ocrEng.Dispose();
}
/// <summary>
/// Processes a specified region in the image using the specified page layout analysis mode.
/// </summary>
/// <remarks>
/// You can only have one result iterator open at any one time.
/// </remarks>
/// <param name="image">The image to process.</param>
/// <param name="region">The image region to process.</param>
/// <param name="pageSegMode">The page layout analyasis method to use.</param>
/// <returns>A result iterator</returns>
public Page Process(Pix image, Rect region, PageSegMode? pageSegMode = null)
{
return Process(image, null, new Rect(0, 0, image.Width, image.Height), pageSegMode);
}
internal static extern void TessBaseAPISetPageSegMode(IntPtr ocr, PageSegMode mode);
public void set_value(PageSegMode mode)
{
this.mode = mode;
}
/// <summary>
/// Process the specified bitmap image.
/// </summary>
/// <remarks>
/// Please consider <see cref="TesseractEngine.Process(Pix, String, Rect, PageSegMode?)"/> instead. This is because
/// this method must convert the bitmap to a pix for processing which will add additional overhead.
/// Leptonica also supports a large number of image pre-processing functions as well.
/// </remarks>
/// <param name="image">The image to process.</param>
/// <param name="inputName">Sets the input file's name, only needed for training or loading a uzn file.</param>
/// <param name="region">The region of the image to process.</param>
/// <param name="pageSegMode">The page segmentation mode.</param>
/// <returns></returns>
public Page Process(Bitmap image, string inputName, Rect region, PageSegMode? pageSegMode = null)
{
var pix = PixConverter.ToPix(image);
var page = Process(pix, inputName, region, pageSegMode);
new PageDisposalHandle(page, pix);
return page;
}
