资源介绍
http://blog.****.net/xiaoxiao108/article/details/7226120
前些天，12306这个网站挺火的，看到网上出现了各种各样的登陆、订票插件跟程序。虽然没经历过春运，看到网上各种各样的插件跟工具挺有意思的，下载了几个看了看，都挺不错的。印象中有个java版本的订票程序里面有个验证码识别功能，用tesseract-ocr来识别验证码的，如果验证码不是很复杂识别效果还可以。
开发环境 vs2008
开发语言C#
使用方法很简单
1.下载tesseract 的.net 类库tessnet2_32.dll ，添加引用。 http://www.pixel-technology.com/freeware/tessnet2/
2.下载tesseract 相对应的语言包。 http://code.google.com/p/tesseract-ocr/downloads/list
3.调用tesseract 的方法进行识别。
具体代码
1.读取网上的验证码到pictureBox中
//string url = "https://dynamic.12306.cn/otsweb/passCodeAction.do?rand=lrand";
string url = "http://static.baixing.net/pages/mobile.php?c=bcqsFelX+vKQcrnIbhyDYQ==/2.jpg";
HttpWebRequest request = WebRequest.Create(url) as HttpWebRequest;
HttpWebResponse response = request.GetResponse() as HttpWebResponse;
System.IO.Stream responseStream = response.GetResponseStream();
this.pictureBox1.Image = Image.FromStream(responseStream);
2.OCR类
public class Ocr
{
public void DumpResult(List result)
{
foreach (tessnet2.Word word in result)
//Console.WriteLine("{0} : {1}", word.Confidence, word.Text);
MessageBox.Show(string.Format("{0} : {1}", word.Confidence, word.Text));
}
public List DoOCRNormal(Bitmap image, string lang)
{
tessnet2.Tesseract ocr = new tessnet2.Tesseract();
ocr.Init(null, lang, false);
List result = ocr.DoOCR(image, Rectangle.Empty);
DumpResult(result);
return result;
}
System.Threading.ManualResetEvent m_event;
public void DoOCRMultiThred(Bitmap image, string lang)
{
tessnet2.Tesseract ocr = new tessnet2.Tesseract();
ocr.Init(null, lang, false);
// If the OcrDone delegate is not null then this'll be the multithreaded version
ocr.OcrDone = new tessnet2.Tesseract.OcrDoneHandler(Finished);
// For event to work, must use the multithreaded version
ocr.ProgressEvent += new tessnet2.Tesseract.ProgressHandler(ocr_ProgressEvent);
m_event = new ManualResetEvent(false);
ocr.DoOCR(image, Rectangle.Empty);
// Wait here it's finished
m_event.WaitOne();
}
public void Finished(List result)
{
DumpResult(result);
m_event.Set();
}
void ocr_ProgressEvent(int percent)
{
Console.WriteLine("{0}% progression", percent);
}
}
3.调用ocr类的方法识别验证码
Ocr ocr = new Ocr();
//using (Bitmap bmp = new Bitmap(@"D:\temp\ocr\b1.bmp"))
using (Bitmap bmp = new Bitmap(this.pictureBox1.Image))
{
tessnet2.Tesseract tessocr = new tessnet2.Tesseract();
tessocr.SetVariable("tessedit_char_whitelist", "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ");
tessocr.Init(null, "eng", false);
tessocr.GetThresholdedImage(bmp, Rectangle.Empty).Save(System.Environment.GetEnvironmentVariable("TEMP") + "\\" + Guid.NewGuid().ToString() + ".bmp");
// Tessdata directory must be in the directory than this exe
//Console.WriteLine("Multithread version");
//ocr.DoOCRMultiThred(bmp, "eng");
//Console.WriteLine("Normal version");
ocr.DoOCRNormal(bmp, "eng");
}
不足之处，如果验证码比较复杂，识别效果不太好，为了提高验证率，可以自己进行训练。具体可参考http://code.google.com/p/tesseract-ocr/
如果你发现有什么不合理的，需要改进的地方，或者你有什么更好的实现方法邮件联系328452421@qq.com（qq常年不在线，邮件联系） 朱晓 。相互交流 谢谢
http://blog.****.net/xiaoxiao108/article/details/7226120