Procedure for tesseract-ocr3.02 Character Recognition

1. Slave;

2. store all the files in the E: \ OCR \ tesseract_ocr3.02 folder and decompress them;

3. Open the Tesseract. sln project under the tesseract-ocr-3.02-vs2008 folder;

4. Copy the corresponding files under the tesseract-ocr-3.02.02 folder such as API, ccmain to E: \ OCR \ tesseract_ocr3.02 \ tesseract-ocr-3.02-vs2008 \ Tesseract-OCR folder, copy the include folder under the leptonica-1.68-win32-lib-include-dirs folder to E: \ OCR \ tesseract_ocr3.02 \ tesseract-ocr-3.02-vs2008 folder, copy files under the tesseract_ocr3.02 \ tesseract-ocr-3.02.02 \ Tesseract-OCR \ vs2008 \ port folder to E: \ OCR \ tesseract_ocr3.02 \ tesseract-ocr-3.02-vs2008 \ Tesseract-OCR \ vs2008 \ port folder, copy the Lib folder under the leptonica-1.68-win32-lib-include-dirs folder to the tesseract_ocr3.02 \ tesseract-ocr-3.02-vs2008 folder;

5. Run equationdetect in the ccmain folder. static const string kcharstoex [] = {"'","'","\"","\\",",",". ",
"<", ">", "," "}; Modify it to static const string kcharstoex [] = {"'","'","\"","\\",",",". "," <","> "," <",">, this file does not exist in Version 3.01, and compilation 3.01 does not need to make any changes to the source file)

6. recompile the entire solution;

7. Create a blank project in the console. The configuration environment is similar to the Tesseract project. Create a tessdata folder under the execution folder to store the chi_sim.traineddata file. The sample code is as follows:

 

#include "allheaders.h"#include "baseapi.h"#include "basedir.h"#include "strngs.h"#include "tesseractmain.h"#include "tprintf.h"int main(int argc, char **argv){  tesseract::TessBaseAPI api;  STRING tessdata_dir;  truncate_path(argv[0], &tessdata_dir);  int rc = api.Init(tessdata_dir.string(), NULL);  if (rc)   {    fprintf(stderr, ("Could not initialize tesseract.\n"));    exit(1);  }  api.End();  // Make the order of args a bit more forgiving than it used to be.  const char* lang = "chi_sim";//eng  const char* image = "E:\\OCR\\tesseract_ocr3.02\\tesseract-ocr-3.02-vs2008\\tesseract-ocr\\vs2008\\Debug\\ABC.tif";//NULL;  const char* output = "E:\\OCR\\tesseract_ocr3.02\\tesseract-ocr-3.02-vs2008\\tesseract-ocr\\vs2008\\Debug\\xxxxx";//NULL;    tesseract::PageSegMode pagesegmode = tesseract::PSM_AUTO;  int arg = 1;  api.SetOutputName(output);  rc = api.Init(tessdata_dir.string(), lang, tesseract::OEM_DEFAULT,                &(argv[arg]), argc - arg, NULL, NULL, false);  if (rc)  {    fprintf(stderr, ("Could not initialize tesseract.\n"));    exit(1);  }  if (api.GetPageSegMode() == tesseract::PSM_SINGLE_BLOCK)  {    api.SetPageSegMode(pagesegmode);  }  tprintf("Tesseract Open Source OCR Engine v%s with Leptonica\n", tesseract::TessBaseAPI::Version());  FILE* fin = fopen(image, "rb");  if (fin == NULL)   {    fprintf(stderr, ("Cannot open input file: %s\n"), image);    exit(2);  }  fclose(fin);  PIX   *pixs;  if ((pixs = pixRead(image)) == NULL)  {    fprintf(stderr, ("Unsupported image type.\n"));    exit(3);  }  pixDestroy(&pixs);  STRING text_out;  if (!api.ProcessPages(image, NULL, 0, &text_out))   {    fprintf(stderr, ("Error during processing.\n"));  }  bool output_hocr = false;  api.GetBoolVariable("tessedit_create_hocr", &output_hocr);  bool output_box = false;  api.GetBoolVariable("tessedit_create_boxfile", &output_box);  STRING outfile = output;  outfile += output_hocr ? ".html" : output_box ? ".box" : ".txt";  FILE* fout = fopen(outfile.string(), "wb");  if (fout == NULL)   {    fprintf(stderr, ("Cannot create output file %s\n"), outfile.string());    exit(1);  }  fwrite(text_out.string(), 1, text_out.length(), fout);  fclose(fout);  return 0;                      // Normal exit}

 

The code above is compiled and run successfully. You can directly enter the entire image for character recognition. The effect is average.