Question

我可以通过命令提示符使用新训练的tessedata（版本3.02）获得正确的OCR输出，但我希望在带有DLL ref的C＃代码中使用相同的输出。我已尝试使用tessnet2_32.dll引用但是它抛出了异常所以如何使用或访问tesseract 3.02版本训练的tessedata使用DLL引用通过C＃代码？

Answer 1

To access or use tesseract 3.02 trained data we have to create separate wrapper class like below.

using System;
using System.IO;
using System.Diagnostics;
using System.Drawing;

/// <summary>
/// Summary description for TesseractOCR
/// </summary>
/// 
namespace tesseractThree
{
    public class TesseractOCR
    {
        public TesseractOCR()
        {
            //
            // TODO: Add constructor logic here
            //
        }

        private string commandpath;
        private string outpath;
        private string tmppath;


        public TesseractOCR(string commandpath)
        {
            this.commandpath = commandpath;
            tmppath = System.Environment.GetFolderPath(Environment.SpecialFolder.ApplicationData) + @"\out.tif";
            outpath = System.Environment.GetFolderPath(Environment.SpecialFolder.ApplicationData) + @"\out.txt";
        }

        public string analyze(string filename,string lang,bool noLine)
        {
            string args = filename + " " + outpath.Replace(".txt", "");
            ProcessStartInfo startinfo;
            if (noLine == true)
            {
                startinfo = new ProcessStartInfo(commandpath, args + " -l " + lang + " -psm 6");
            }
            else
            {
                startinfo = new ProcessStartInfo(commandpath, args + " -l " + lang);
            }
            startinfo.CreateNoWindow = true;
            startinfo.UseShellExecute = false;
            Process.Start(startinfo).WaitForExit();

            string ret = "";
            using (StreamReader r = new StreamReader(outpath))
            {
                string content = r.ReadToEnd();
                ret = content;
            }
            File.Delete(outpath);
            return ret;
        }

        public string OCRFromBitmap(Bitmap bmp,string lang,bool noLine)
        {

            bmp.Save(tmppath, System.Drawing.Imaging.ImageFormat.Tiff);
            string ret = analyze(tmppath,lang,noLine);
            File.Delete(tmppath);
            return ret;
        }
      /*  public string OCRFromFile(string filename)
        {
            return analyze(filename);
        }*/
    }
}

    //Usage of this class
                string lang = "enc";
                Bitmap b = new Bitmap(@"D:\Image\enc.test_font.exp0.tif");
                TesseractOCR ocr = new TesseractOCR(@"C:\Program Files\Tesseract-OCR\tesseract.exe");
                string result = ocr.OCRFromBitmap(b, lang,true);
                Label1.Text = result;

OR Refer below link for more details.
https://gist.github.com/yatt/915443

Answer 2

那是针对Tesseract 2.04的。您需要一个与3.02版本兼容的.NET wrapper。

Answer 3

使用tesseractengine3.dll我们可以使用tesseract v3.02训练数据，如下所示。

using System;
using System.Collections.Generic;
using System.Linq;
using System.Web;
using System.Web.UI;
using System.Web.UI.WebControls;
using tesseract;
using System.Drawing;
using System.IO;

public enum TesseractEngineMode : int
{
    /// <summary>
    /// Run Tesseract only - fastest
    /// </summary>
    TESSERACT_ONLY = 0,

    /// <summary>
    /// Run Cube only - better accuracy, but slower
    /// </summary>
    CUBE_ONLY = 1,

    /// <summary>
    /// Run both and combine results - best accuracy
    /// </summary>
    TESSERACT_CUBE_COMBINED = 2,

    /// <summary>
    /// Specify this mode when calling init_*(),
    /// to indicate that any of the above modes
    /// should be automatically inferred from the
    /// variables in the language-specific config,
    /// command-line configs, or if not specified
    /// in any of the above should be set to the
    /// default OEM_TESSERACT_ONLY.
    /// </summary>
    DEFAULT = 3
}

public enum TesseractPageSegMode : int
{
    /// <summary>
    /// Fully automatic page segmentation
    /// </summary>
    PSM_AUTO = 0,

    /// <summary>
    /// Assume a single column of text of variable sizes
    /// </summary>
    PSM_SINGLE_COLUMN = 1,

    /// <summary>
    /// Assume a single uniform block of text (Default)
    /// </summary>
    PSM_SINGLE_BLOCK = 2,

    /// <summary>
    /// Treat the image as a single text line
    /// </summary>
    PSM_SINGLE_LINE = 3,

    /// <summary>
    /// Treat the image as a single word
    /// </summary>
    PSM_SINGLE_WORD = 4,

    /// <summary>
    /// Treat the image as a single character
    /// </summary>
    PSM_SINGLE_CHAR = 5
}

public partial class importDLL : System.Web.UI.Page
{

    private TesseractProcessor m_tesseract = null;

    //private const string m_path = @"..\..\data\";
    private const string m_path = @"D:\tessdata-3.02\";
    private const string m_lang = "eng";


    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);
    }


}

如何在C＃中使用tesseract 3.02训练数据？

3 个答案: