public TesseractOCREngine(Image image)
{
this.image = image;
m_tesseract = new TesseractProcessor();
bool succeed = m_tesseract.Init(m_path, m_lang, (int)TesseractEngineMode.DEFAULT);
m_tesseract.SetVariable("tessedit_char_whitelist", "0123456789-.eExXpP/\\");
m_tesseract.SetVariable("tessedit_pageseg_mode", ((int)TesseractPageSegMode.PSM_SINGLE_LINE).ToString());
System.Environment.CurrentDirectory = System.IO.Path.GetFullPath(m_path);
}
public Form1()
{
InitializeComponent();
m_tesseract = new TesseractProcessor();
m_tesseract.Init(m_path, m_lang, (int)TesseractEngineMode.DEFAULT);
m_tesseract.SetVariable("tessedit_pageseg_mode", TesseractPageSegMode.PSM_AUTO.ToString());
}
private const string m_lang = "eng"; // 识别的语言文件名称
public string read2dig(Image image)
{
TesseractProcessor m_tesseract = new TesseractProcessor();
;
bool succeed = m_tesseract.Init(m_path, m_lang, (int)TesseractEngineMode.DEFAULT);
if (!succeed)
{
return("tesseract初始化失败");
}
m_tesseract.SetVariable("tessedit_char_whitelist", "0123456789"); //设置识别变量，当前只能识别数字。
image.Save("./img/原图.bmp", System.Drawing.Imaging.ImageFormat.Bmp);
Bitmap bmp = ToGray(image);
bmp.Save("./img/灰度图.bmp", System.Drawing.Imaging.ImageFormat.Bmp);
bmp = ConvertTo1Bpp1(bmp, 110);
bmp.Save("./img/二值图.bmp", System.Drawing.Imaging.ImageFormat.Bmp);
bmp = Zoom(bmp, 0.25);
bmp.Save("./img/放大缩小后.bmp", System.Drawing.Imaging.ImageFormat.Bmp);
m_tesseract.Clear();
m_tesseract.ClearAdaptiveClassifier();
return(m_tesseract.Apply(bmp));
}
public static string Recognize(string url)
{
WebRequest request = WebRequest.Create(url);
WebResponse response = request.GetResponse();
Stream st = response.GetResponseStream();
if (st == null)
{
return(string.Empty);
}
try
{
Bitmap bitmap = ((Bitmap)Bitmap.FromStream(st)).toGray().biLinear(2);
//var newBitmap= KiCut(bitmap, 12, 22, bitmap.Width - 12, 22);
string tessdata = Environment.CurrentDirectory + "\\tessdata\\";
string language = "eng";//设置训练文件的名称，后缀traineddata之前的名称
int oem = 3;
// http://www.lixin.me/blog/2012/05/26/29536 训练
using (TesseractProcessor processor = new TesseractProcessor())
{
//初始化
bool initFlage = processor.Init(tessdata, language, oem);
//processor.GetTesseractEngineVersion();//获取版本号
/*
* Set the current page segmentation mode. Defaults to PSM_SINGLE_BLOCK.
* The mode is stored as an IntParam so it can also be modified by ReadConfigFile or SetVariable("tessedit_pageseg_mode", mode as string). */
//processor.SetPageSegMode(ePageSegMode.PSM_SINGLE_BLOCK);
//设置ROI（图像的感兴趣区域）
processor.UseROI = true;
processor.ROI = new Rectangle(24, 0, bitmap.Width - 24, bitmap.Height);
//设置识别的变量 如果是自定义培训的文件 可以不用设置
//必须在初始化后调用
processor.SetVariable("tessedit_char_whitelist", "0123456789.");
//processor.SetVariable("tessedit_thresholding_method", "1"); //图像处理阀值是否打开
//processor.SetVariable("save_best_choices", "T");
using (Bitmap bmp = bitmap)
{
int i = 3;
oem = i;
string text = processor.Recognize(bmp);
char[] charsToTrim = { '\\', 'n', '\\', 'n' };
return(text.TrimEnd(charsToTrim));
//Console.WriteLine(
// string.Format(
// "RecognizeMode: {1}\nText:\n{0}\n++++++\n", text, ((eOcrEngineMode)oem).ToString()));
}
}
}
catch (Exception)
{
return(string.Empty);
}
}
private void OCR_Load(object sender, EventArgs e)
{
m_tesseract = new TesseractProcessor();
bool succeed = m_tesseract.Init(m_path, m_lang, 3);
if (!succeed)
{
MessageBox.Show("Tesseract initialization failed. The application will exit.");
Application.Exit();
}
m_tesseract.SetVariable("tessedit_char_whitelist", "ABCDEFHKLMNPQRSTUVXYZ-.1234567890").ToString(); //Training Sets //abcdefghijklmnopqrstuvxyzABCDEFGHIJKLMNOPQRSTUVXYZ-.1234567890
System.Environment.CurrentDirectory = System.IO.Path.GetFullPath(m_path);
}
private void MainForm_Load(object sender, EventArgs e)
{
int soluong;
int tong;
tong = int.Parse(txt_tong.Text);
soluong = int.Parse(bss.so_luong_xe());
txt_conlai.Text = (tong - soluong).ToString();
dataGridView1.DataSource = bss.Danhsachxe();
capture = new Emgu.CV.Capture();
timer1.Enabled = true;
full_tesseract = new TesseractProcessor();
bool succeed = full_tesseract.Init(m_path, m_lang, 3);
if (!succeed)
{
MessageBox.Show("Tesseract initialization failed. The application will exit.");
Application.Exit();
}
full_tesseract.SetVariable("tessedit_char_whitelist", "ABCDEFHKLMNPRSTVXY1234567890").ToString();
ch_tesseract = new TesseractProcessor();
succeed = ch_tesseract.Init(m_path, m_lang, 3);
if (!succeed)
{
MessageBox.Show("Tesseract initialization failed. The application will exit.");
Application.Exit();
}
ch_tesseract.SetVariable("tessedit_char_whitelist", "ABCDEFHKLMNPRSTUVXY").ToString();
num_tesseract = new TesseractProcessor();
succeed = num_tesseract.Init(m_path, m_lang, 3);
if (!succeed)
{
MessageBox.Show("Tesseract initialization failed. The application will exit.");
Application.Exit();
}
num_tesseract.SetVariable("tessedit_char_whitelist", "1234567890").ToString();
m_path = System.Environment.CurrentDirectory + "\\";
string[] ports = SerialPort.GetPortNames();
for (int i = 0; i < box.Length; i++)
{
box[i] = new PictureBox();
}
}
void CheckInPage_Loaded(object sender, RoutedEventArgs e)
{
full_tesseract = new TesseractProcessor();
bool succeed = full_tesseract.Init(m_path, m_lang, 3);
if (!succeed)
{
System.Windows.MessageBox.Show("Lỗi thư viện Tesseract. Chương trình cần kết thúc.");
}
full_tesseract.SetVariable("tessedit_char_whitelist", "ACDFHKLMNPRSTVXY1234567890").ToString();
ch_tesseract = new TesseractProcessor();
succeed = ch_tesseract.Init(m_path, m_lang, 3);
if (!succeed)
{
System.Windows.MessageBox.Show("Lỗi thư viện Tesseract. Chương trình cần kết thúc.");
}
ch_tesseract.SetVariable("tessedit_char_whitelist", "ACDEFHKLMNPRSTUVXY").ToString();
num_tesseract = new TesseractProcessor();
succeed = num_tesseract.Init(m_path, m_lang, 3);
if (!succeed)
{
System.Windows.MessageBox.Show("Lỗi thư viện Tesseract. Chương trình cần kết thúc.");
}
num_tesseract.SetVariable("tessedit_char_whitelist", "1234567890").ToString();
System.Environment.CurrentDirectory = System.IO.Path.GetFullPath(m_path);
for (int i = 0; i < box.Length; i++)
{
box[i] = new PictureBox();
}
string folder = System.AppDomain.CurrentDomain.BaseDirectory + "\\ImageTest";
foreach (string fileName in Directory.GetFiles(folder, "*.bmp", SearchOption.TopDirectoryOnly))
{
lstimages.Add(System.IO.Path.GetFullPath(fileName));
}
foreach (string fileName in Directory.GetFiles(folder, "*.jpg", SearchOption.TopDirectoryOnly))
{
lstimages.Add(System.IO.Path.GetFullPath(fileName));
}
pic1 = (hostPic1.Child as System.Windows.Forms.PictureBox);
pic2 = (hostPic2.Child as System.Windows.Forms.PictureBox);
img1 = (hostImg1.Child as Emgu.CV.UI.ImageBox);
}
//private string _Caption = "http://www.laptrinhvb.net";
#endregion
private void FrmMain_Load(object sender, EventArgs e)
{
full_tesseract = new TesseractProcessor();
bool succeed = full_tesseract.Init(m_path, m_lang, 3);
if (!succeed)
{
MessageBox.Show("Lỗi thư viện Tesseract. Chương trình cần kết thúc.");
Application.Exit();
}
full_tesseract.SetVariable("tessedit_char_whitelist", "ACDFHKLMNPRSTVXY1234567890").ToString();
ch_tesseract = new TesseractProcessor();
succeed = ch_tesseract.Init(m_path, m_lang, 3);
if (!succeed)
{
MessageBox.Show("Lỗi thư viện Tesseract. Chương trình cần kết thúc.");
Application.Exit();
}
ch_tesseract.SetVariable("tessedit_char_whitelist", "ACDEFHKLMNPRSTUVXY").ToString();
num_tesseract = new TesseractProcessor();
succeed = num_tesseract.Init(m_path, m_lang, 3);
if (!succeed)
{
MessageBox.Show("Lỗi thư viện Tesseract. Chương trình cần kết thúc.");
Application.Exit();
}
num_tesseract.SetVariable("tessedit_char_whitelist", "1234567890").ToString();
System.Environment.CurrentDirectory = System.IO.Path.GetFullPath(m_path);
for (int i = 0; i < box.Length; i++)
{
box[i] = new PictureBox();
}
string folder = Application.StartupPath + "\\ImageTest";
foreach (string fileName in Directory.GetFiles(folder, "*.bmp", SearchOption.TopDirectoryOnly))
{
lstimages.Add(Path.GetFullPath(fileName));
}
foreach (string fileName in Directory.GetFiles(folder, "*.jpg", SearchOption.TopDirectoryOnly))
{
lstimages.Add(Path.GetFullPath(fileName));
}
}
/* private void button3_Click(object sender, EventArgs e)
* {
* if (capture != null)
* {
* timer1.Enabled = false;
* pictureBox_XeRA.Image = null;
* IF.pictureBox2.Image = null;
* capture.QueryFrame().Save("aa.bmp");
* FileStream fs = new FileStream(m_path + "aa.bmp", FileMode.Open, FileAccess.Read);
* Image temp = Image.FromStream(fs);
* fs.Close();
* pictureBox_XeRA.Image = temp;
* IF.pictureBox2.Image = temp;
* pictureBox_XeRA.Update();
* IF.pictureBox2.Update();
* Image temp1;
* string temp2, temp3;
* /*Reconize(m_path + "aa.bmp", out temp1, out temp2, out temp3);
* pictureBox_XeVAO.Image = temp1;
* if(temp3 == "")
* text_BiensoVAO.Text = "ko nhận dạng dc biển số";
* else
* text_BiensoVAO.Text = temp3;
*
* timer1.Enabled = true;
* }
*
* }*/
private void MainForm_Load_1(object sender, EventArgs e)
{
capture = new Emgu.CV.Capture();
timer2.Enabled = true;
timer4.Start();
IF = new ImageForm();
full_tesseract = new TesseractProcessor();
bool succeed = full_tesseract.Init(m_path, m_lang, 3);
if (!succeed)
{
MessageBox.Show("Tesseract initialization failed. The application will exit.");
Application.Exit();
}
full_tesseract.SetVariable("tessedit_char_whitelist", "ABCDEFHKLMNPRSTVXY1234567890").ToString();
ch_tesseract = new TesseractProcessor();
succeed = ch_tesseract.Init(m_path, m_lang, 3);
if (!succeed)
{
MessageBox.Show("Tesseract initialization failed. The application will exit.");
Application.Exit();
}
ch_tesseract.SetVariable("tessedit_char_whitelist", "ABCDEFHKLMNPRSTUVXY").ToString();
num_tesseract = new TesseractProcessor();
succeed = num_tesseract.Init(m_path, m_lang, 3);
if (!succeed)
{
MessageBox.Show("Tesseract initialization failed. The application will exit.");
Application.Exit();
}
num_tesseract.SetVariable("tessedit_char_whitelist", "1234567890").ToString();
m_path = System.Environment.CurrentDirectory + "\\";
string[] ports = SerialPort.GetPortNames();
for (int i = 0; i < box.Length; i++)
{
box[i] = new PictureBox();
}
}
protected void Page_Load(object sender, EventArgs e)
{
var image = System.Drawing.Image.FromFile(@"D:\Image\Capture1T.tif");
m_tesseract = new TesseractProcessor();
bool succeed = m_tesseract.Init(m_path, m_lang, (int)TesseractEngineMode.DEFAULT);
if (!succeed)
{
}
m_tesseract.SetVariable("tessedit_pageseg_mode", ((int)TesseractPageSegMode.PSM_SINGLE_LINE).ToString());
m_tesseract.Clear();
m_tesseract.ClearAdaptiveClassifier();
string outValue = m_tesseract.Apply(image);
Response.Write(outValue);
}
static oTesseractRequest __ocrExecute(oTesseractRequest req, Bitmap bitmap)
{
using (TesseractProcessor processor = new TesseractProcessor())
{
// call SetVariable() method before passing image(api->SetImage(image))
// 0: otsu
// 1: isodata local adaptive
// 2: sauvola local adaptive => not implement yet
//processor.SetVariable("tessedit_thresholding_method", "0");
processor.SetVariable("tessedit_thresholding_method", "1");
//processor.InitForAnalysePage();
//processor.SetPageSegMode(ePageSegMode.PSM_AUTO_ONLY);
//var success = processor.Init(req.data_path, req.lang, (int)eOcrEngineMode.OEM_DEFAULT);
//var imageColor = new Emgu.CV.Mat();
//var imageCV = new Emgu.CV.Image<Emgu.CV.Structure.Bgr, byte>(bitmap); //Image Class from Emgu.CV
////var imageCV = new Emgu.CV.Image<Emgu.CV.Structure.Gray, byte>(bitmap); //Image Class from Emgu.CV
//var image = imageCV.Mat; //This is your Image converted to Mat
//if (image.NumberOfChannels == 1)
// Emgu.CV.CvInvoke.CvtColor(image, imageColor, Emgu.CV.CvEnum.ColorConversion.Gray2Bgr);
//else
// image.CopyTo(imageColor);
using (var m0 = new MemoryStream())
{
bitmap.Save(m0, ImageFormat.Jpeg);
//using (Bitmap bmp = Bitmap.FromFile(@"C:\temp\1.jpg") as Bitmap)
using (Bitmap bmp = new Bitmap(m0))
{
DocumentLayout doc = null;
switch (req.command)
{
case TESSERACT_COMMAND.GET_TEXT:
//string s = tes.GetText().Trim();
//req.output_text = s;
//req.output_count = s.Length;
req.ok = 1;
break;
default:
unsafe
{
doc = processor.AnalyseLayout(bmp);
}
if (doc != null)
{
var bs = new List <string>();
if (doc.Blocks.Count > 0)
{
for (int i = 0; i < doc.Blocks.Count; i++)
{
for (int j = 0; j < doc.Blocks[i].Paragraphs.Count; j++)
{
bs.AddRange(doc.Blocks[j].Paragraphs[j].Lines
.Select(x => string.Format(
"{0}_{1}_{2}_{3}", x.Left, x.Top, x.Right, x.Bottom)));
}
}
}
req.output_format = "left_top_right_bottom";
req.output_text = string.Join("|", bs.ToArray());
req.output_count = bs.Count;
req.ok = 1;
}
break;
}
}
}
}
return(req);
}
public static string Recognize(string url,Rectangle rectangle)
{
WebRequest request = WebRequest.Create(url);
WebResponse response = request.GetResponse();
Stream st = response.GetResponseStream();
if (st == null)
{
return string.Empty;
}
Bitmap bitmap = ((Bitmap)Bitmap.FromStream(st)).toGray().biLinear(2);
//var newBitmap= KiCut(bitmap, 12, 22, bitmap.Width - 12, 22);
string tessdata = Environment.CurrentDirectory + "\\tessdata\\";
string language = "eng";//设置训练文件的名称，后缀traineddata之前的名称
int oem = 3;
// http://www.lixin.me/blog/2012/05/26/29536 训练
using (TesseractProcessor processor = new TesseractProcessor())
{
//初始化
bool initFlage = processor.Init(tessdata, language, oem);
//processor.GetTesseractEngineVersion();//获取版本号
/*
Set the current page segmentation mode. Defaults to PSM_SINGLE_BLOCK.
The mode is stored as an IntParam so it can also be modified by ReadConfigFile or SetVariable("tessedit_pageseg_mode", mode as string). */
//processor.SetPageSegMode(ePageSegMode.PSM_SINGLE_BLOCK);
//设置ROI（图像的感兴趣区域）
if (rectangle != null)
{
processor.UseROI = true;
processor.ROI = rectangle;
}
//设置识别的变量 如果是自定义培训的文件 可以不用设置
//必须在初始化后调用
processor.SetVariable("tessedit_char_whitelist", "0123456789.");
//processor.SetVariable("tessedit_thresholding_method", "1"); //图像处理阀值是否打开
//processor.SetVariable("save_best_choices", "T");
using (Bitmap bmp = bitmap)
{
string text = processor.Recognize(bmp);
char[] charsToTrim = { '\\', 'n', '\\', 'n' };
return text.TrimEnd(charsToTrim);
}
}
}
public Form1()
{
InitializeComponent();
this.PSN = new Person();
this.imagePath = Application.StartupPath + "/image";
if (Directory.Exists(this.imagePath) == false) Directory.CreateDirectory(this.imagePath);
this.imagePath += "/";
Init();
this.F = new mPublic();
this.CFG = new Config();
this.MAPP = new Mapping();
this.RMap = CFG.rect_map_now;
this.thKill = new Thread(new ThreadStart(AutoKill));
this.stopKill = true;
this.ocrNumber = new OCR();
/*
//初始化数字识别引擎
ocrNumber = new TesseractProcessor();
bInitN = ocrNumber.Init(@".\", "chi_sim", 3);
if (!bInitN) { MessageBox.Show("智能识别引擎初始化失败！"); return; }
ocrNumber.SetVariable("tessedit_char_whitelist", ".,2134567890 ");
* */
//初始化汉字引擎
ocrText = new TesseractProcessor();
bInitT = ocrText.Init(@".\", "chi_sim", 3);
if (!bInitT) { MessageBox.Show("智能识别引擎初始化失败！"); return; }
ocrText.SetVariable("tessedit_char_blacklist", "fi蜘fi'£郜QWERTYUIOPASDFGHJKLZXCVBNM1234567890μqwertyuiopasdfghjklzxcvbnm");
}
/*
private string GetNumber(Bitmap image)
{
UnCodebase ud = new UnCodebase(image);
ud.DealMap2();
image = ud.bmpobj;
pictureBox1.Image = image;
if (!bInitT)
{
ocrNumber = new TesseractProcessor();
bool inited = ocrNumber.Init(@".\", "chi_sim", 3);
if (!inited) { MessageBox.Show("智能识别引擎初始化失败！"); return ""; }
ocrNumber.SetVariable("tessedit_char_whitelist", ".,3214567890 ");
}
ocrNumber.Clear();
ocrNumber.ClearAdaptiveClassifier();
string result = ocrNumber.Recognize(image);
return result;
}
* */
private string GetChiSim(Bitmap image)
{
UnCodebase ud = new UnCodebase(image);
ud.DealText();
image = ud.image;
//pictureBox1.Image = image; textBox3.Text = ud.txt;
if (!bInitT)
{
ocrText = new TesseractProcessor();
bInitT = ocrText.Init(@".\", "chi_sim", 3);
if (!bInitT) { MessageBox.Show("智能识别引擎初始化失败！"); return ""; }
ocrText.SetVariable("tessedit_char_blacklist", "1234567890μqwertyuiopasdfghjklzxcvbnm'I flU]M.H flO");
}
ocrText.Clear();
ocrText.ClearAdaptiveClassifier();
string result = ocrText.Recognize(image);
return result;
}
private void DoOCR(IImage image)
{
ClearRenderingData();
string variable = "tessedit_pageseg_mode";
int storedOSD = 0;
_ocrProcessor.GetIntVariable(variable, ref storedOSD);
try
{
// Fully automatic page segmentation
int fully_psm_auto = 3;
_ocrProcessor.SetVariable(variable, fully_psm_auto.ToString());
///// DEMO ONLY
using (Image bitmap = ToImage(image))
{
///// DEMO ONLY
_ocrProcessor.Clear();
_ocrProcessor.ClearAdaptiveClassifier();
BlockList blocks = _ocrProcessor.DetectBlocks(bitmap);
this.UpdateImageViewer(blocks);
///// DEMO ONLY
_ocrProcessor.Clear();
_ocrProcessor.ClearAdaptiveClassifier();
string result = _ocrProcessor.Apply(bitmap);
List <Word> detectedWords = _ocrProcessor.RetriveResultDetail();
this.UpdateImageViewer(detectedWords);
if (!string.IsNullOrEmpty(result))
{
MessageBox.Show(result);
}
}
}
catch
{
}
finally
{
_ocrProcessor.SetVariable(variable, storedOSD.ToString());
}
//_ocrProcessor.Clear();
//_ocrProcessor.ClearAdaptiveClassifier();
//using (Image bitmap = ToImage(image))
//{
// //string result = _ocrProcessor.Apply(_fileName);
// string result = _ocrProcessor.Apply(bitmap);
// List<Word> detectedWords = _ocrProcessor.RetriveResultDetail();
// this.UpdateImageViewer(detectedWords);
// if (!string.IsNullOrEmpty(result))
// MessageBox.Show(result);
//}
}
private void button2_Click(object sender, EventArgs e)
{
Image mg = Image.FromFile(Application.StartupPath + "\\1.jpg");
////Image mg = Image.FromFile("e:\\test24.tif");
imageViewer1.Initialize(new OCRRenderingData(), new OCRAnalysisRender(imageViewer1));
imageViewer1.Image = mg;
TesseractProcessor _ocrProcessor = null;
_ocrProcessor = new TesseractProcessor();
_ocrProcessor.UseROI = false;
imageViewer1.RectA = new Rectangle(0, 0, 0, 0);
//string _tessData = "E:\\tesseract\\tesseract-ocr-dotnet-3.01\\tessdata\\";
string _tessData = Application.StartupPath + "\\tessdata\\";
string _lang = "eng";
int _ocrEngineMode = 3;
bool status = _ocrProcessor.Init(_tessData, _lang, _ocrEngineMode);
Console.WriteLine(string.Format("[DEBUG] Init status: {0}", status));
//string text = _ocrProcessor.Recognize(mg);
//MessageBox.Show(text);
////**************************************************** converts 1bpp to 24bpp
//Bitmap image = new Bitmap(@"e:\\test11.tif");
//Bitmap bitmap = new Bitmap(image.Width, image.Height, System.Drawing.Imaging.PixelFormat.Format24bppRgb);
//for (int i = 0; i < image.Width; i++)
//{
// for (int j = 0; j < image.Height; j++)
// {
// Color temp = image.GetPixel(i, j);
// bitmap.SetPixel(i, j, temp);
// }
//}
//bitmap.Save(@"e:\\test24.tif", System.Drawing.Imaging.ImageFormat.Bmp);
//Image mg1 = Image.FromFile("e:\\test24.tif");
////******************************************************************************************************
string variable = "tessedit_pageseg_mode";
// Fully automatic page segmentation
int fully_psm_auto = 3;
_ocrProcessor.SetVariable(variable, fully_psm_auto.ToString());
///// DEMO ONLY
//_ocrProcessor.Clear();
//_ocrProcessor.ClearAdaptiveClassifier();
//BlockList blocks = _ocrProcessor.DetectBlocks(mg);
//this.UpdateImageViewer(blocks);
_ocrProcessor.Clear();
_ocrProcessor.ClearAdaptiveClassifier();
string result = _ocrProcessor.Apply(mg);
List <Word> detectedWords = _ocrProcessor.RetriveResultDetail();
this.UpdateImageViewer(detectedWords);
//MessageBox.Show(result);
button3.Enabled = true;
button4.Enabled = true;
}
