分享一下我老师大神的人工智能教程!零基础,通俗易懂!http://blog.csdn.net/jiangjunshow
也欢迎大家转载本篇文章。分享知识,造福人民,实现我们中华民族伟大复兴!
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";//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}
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。