tesseract-ocr3.02字符识别过程操作步骤

1、  从http://code.google.com/p/tesseract-ocr/downloads/list下载tesseract-ocr-3.02-vs2008、tesseract-ocr-3.02.chi_sim.tar、tesseract-ocr-3.02.02.tar、tesseract-ocr-3.02.02-doc-html.tar、leptonica-1.68-win32-lib-include-dirs相关文件；

2、  将所有文件存放在E:\OCR\tesseract_ocr3.02文件夹下并解压缩；

3、  打开tesseract-ocr-3.02-vs2008文件夹下的tesseract.sln工程；

4、  将tesseract-ocr-3.02.02文件夹下的对应文件如api、ccmain等复制到E:\OCR\tesseract_ocr3.02\tesseract-ocr-3.02-vs2008\tesseract-ocr文件夹下，把leptonica-1.68-win32-lib-include-dirs文件夹下的include文件夹复制到E:\OCR\tesseract_ocr3.02\tesseract-ocr-3.02-vs2008文件夹下，将tesseract_ocr3.02\tesseract-ocr-3.02.02\tesseract-ocr\vs2008\port文件夹下的文件复制到E:\OCR\tesseract_ocr3.02\tesseract-ocr-3.02-vs2008\tesseract-ocr\vs2008\port文件夹下，将leptonica-1.68-win32-lib-include-dirs文件夹下lib文件夹复制到tesseract_ocr3.02\tesseract-ocr-3.02-vs2008文件夹下；

5、  将ccmain文件夹下的equationdetect.cpp文件中的static const STRING kCharsToEx[] = {"'", "`","\"", "\\", ",", ".", "〈", "〉", "《", "》", "」", "「", ""};修改成static const STRING kCharsToEx[] = {"'", "`","\"", "\\", ",",".","<", ">",   "<<",">>",  ""};(注：不改动编译时始终出错，其它方法暂未发现，3.01版本中没有此文件，编译3.01不用对源文件作任何修改)

6、 重新编译整个Solution；

7、 创建一个控制台空工程，配置环境仿照tesseract工程，并在执行文件夹下创建一个tessdata文件夹，里面存放chi_sim.traineddata数据文件，示例代码如下：

#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";//engconst 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
}

以上代码编译运行成功，直接输入整幅图像进行字符识别，效果一般。