1、 從http://code.google.com/p/tesseract-ocr/downloads/list下載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相關文件;
2、 將所有文件存放在E:\OCR\tesseract_ocr3.02文件夾下並解壓縮;
3、 打開tesseract-ocr-3.02-vs2008文件夾下的tesseract.sln工程;
4、 將tesseract-ocr-3.02.02文件夾下的對應文件如api、ccmain等複製到E:\OCR\tesseract_ocr3.02\tesseract-ocr-3.02-vs2008\tesseract-ocr文件夾下,把leptonica-1.68-win32-lib-include-dirs文件夾下的include文件夾複製到E:\OCR\tesseract_ocr3.02\tesseract-ocr-3.02-vs2008文件夾下,將tesseract_ocr3.02\tesseract-ocr-3.02.02\tesseract-ocr\vs2008\port文件夾下的文件複製到E:\OCR\tesseract_ocr3.02\tesseract-ocr-3.02-vs2008\tesseract-ocr\vs2008\port文件夾下,將leptonica-1.68-win32-lib-include-dirs文件夾下lib文件夾複製到tesseract_ocr3.02\tesseract-ocr-3.02-vs2008文件夾下;
5、 將ccmain文件夾下的equationdetect.cpp文件中的static const STRING kCharsToEx[] = {"'", "`","\"", "\\", ",", ".", "〈", "〉", "《", "》", "」", "「", ""};修改成static const STRING kCharsToEx[] = {"'", "`","\"", "\\", ",",".","<", ">", "<<",">>", ""};(注:不改動編譯時始終出錯,其它方法暫未發現,3.01版本中沒有此文件,編譯3.01不用對源文件作任何修改)
6、 重新編譯整個Solution;
7、 創建一個控制檯空工程,配置環境仿照tesseract工程,並在執行文件夾下創建一個tessdata文件夾,裏面存放chi_sim.traineddata數據文件,示例代碼如下:
[cpp] view plaincopy #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* p_w_picpath = "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(p_w_picpath, "rb"); if (fin == NULL) { fprintf(stderr, ("Cannot open input file: %s\n"), p_w_picpath); exit(2); } fclose(fin); PIX *pixs; if ((pixs = pixRead(p_w_picpath)) == NULL) { fprintf(stderr, ("Unsupported p_w_picpath type.\n")); exit(3); } pixDestroy(&pixs); STRING text_out; if (!api.ProcessPages(p_w_picpath, 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 }
以上代碼編譯運行成功,直接輸入整幅圖像進行字符識別,效果一般。