配置EmguCV中的OCR

Posted 2021-02-16 jiceberg420

问题：Unable to create ocr model using Path ‘.‘ and language ‘eng‘.

_ocr = new Tesseract(@".", "eng", OcrEngineMode.TesseractLstmCombined);

纠结了有几个星期，最后终于试出来了。

总结原因有N，如下：

1. 路径写法

path =  @"C:myToolsTess“ 后面需要加上 ，

2. 子文件夹

如果path= @"C:myToolsTess", 那么文件应该放在一个名叫tessdata的子文件夹内，而不是直接放在path的目录下。

3. 版本匹配

刚开始我下载了最新的traindata放在文件夹下面，不好使呀，随后找到了正确的版本，才成功

emgu版本号：3.4.3 所需要的数据版本路径如下

https://github.com/tesseract-ocr/tessdata/blob/4592b8d453889181e01982d22328b5846765eaad/eng.traineddata

https://github.com/tesseract-ocr/tessdata/blob/4592b8d453889181e01982d22328b5846765eaad/osd.traineddata

 

运行通过的代码和文件夹配置如下

string path =  @"C:myToolsTess"; 
 _ocr = new Tesseract(path, "eng", OcrEngineMode.TesseractLstmCombined);

 

参考

从Emgu官网下载的程序包里面的例程。

http://www.emgu.com/wiki/index.php/Main_Page

 

关键代码区域如下

        private void InitOcr(String path, String lang, OcrEngineMode mode)
        {
            try
            {
                if (_ocr != null)
                {
                    _ocr.Dispose();
                    _ocr = null;
                }

                if (String.IsNullOrEmpty(path))
                    path = ".";

                TesseractDownloadLangFile(path, lang);
                TesseractDownloadLangFile(path, "osd"); //script orientation detection
                String pathFinal = path.Length == 0 || path.Substring(path.Length - 1, 1).Equals(Path.DirectorySeparatorChar.ToString())
                    ? path
                    : String.Format("{0}{1}", path, System.IO.Path.DirectorySeparatorChar);

                _ocr = new Tesseract(pathFinal, lang, mode);

            }
            catch (Exception e)
            {
                _ocr = null;
                System.Diagnostics.Debug.Print(e.Message, "Failed to initialize tesseract OCR engine");
            }
        }

        private static void TesseractDownloadLangFile(String folder, String lang)
        {
            String subfolderName = "tessdata";
            String folderName = System.IO.Path.Combine(folder, subfolderName);
            if (!System.IO.Directory.Exists(folderName))
            {
                System.IO.Directory.CreateDirectory(folderName);
            }
            String dest = System.IO.Path.Combine(folderName, String.Format("{0}.traineddata", lang));
            if (!System.IO.File.Exists(dest))
                using (System.Net.WebClient webclient = new System.Net.WebClient())
                {
                    String source =
                        String.Format("https://github.com/tesseract-ocr/tessdata/blob/4592b8d453889181e01982d22328b5846765eaad/{0}.traineddata?raw=true", lang);

                    Console.WriteLine(String.Format("Downloading file from ‘{0}‘ to ‘{1}‘", source, dest));
                    webclient.DownloadFile(source, dest);
                    Console.WriteLine(String.Format("Download completed"));
                }
        }