基于Tesseract-OCR实现的JAVA WEB版OCR(图片转文字)

首先在 Tesseract-OCR官网下载Tesseract-OCR 3.02，以及中文数据包chi_sim.traineddata(简体)

接下来就是新建一个JAVA EE项目，把Tesseract-OCR放在项目WebRoot下。

下面是主要代码：

接受客户端上传过来的图片，使用Tesseract-OCR识别后返回至前台。

[java] view plain copy

print ?

1. package servlet;  
2.   
3. import java.io.IOException;  
4.   
5. import javax.servlet.ServletConfig;  
6. import javax.servlet.ServletException;  
7. import javax.servlet.http.HttpServlet;  
8. import javax.servlet.http.HttpServletRequest;  
9. import javax.servlet.http.HttpServletResponse;  
10.   
11. import util.FileUtil;  
12. import util.OCRUtil;  
13.   
14. import com.jspsmart.upload.File;  
15. import com.jspsmart.upload.SmartUpload;  
16. import com.jspsmart.upload.SmartUploadException;  
17.   
18.   
19. public class OCRServlet extends HttpServlet {  
20.   
21.     public void doPost(HttpServletRequest request, HttpServletResponse response)  
22.             throws ServletException, IOException {  
23.         response.setCharacterEncoding("gbk");  
24.         SmartUpload upload = new SmartUpload();  
25.         ServletConfig sc = this.getServletConfig();  
26.         upload.initialize(sc, request, response);  
27.         File file = null;  
28.         long size = 5*1024*1024;  
29.         upload.setAllowedFilesList("gif,jpg,bmp,png");  
30.         upload.setMaxFileSize(size);  
31.         upload.setCharset("GBK");  
32.         try {  
33.             upload.upload();  
34.             file = upload.getFiles().getFile(0);  
35.             String userPath = "upload\\"+request.getRemoteAddr().replaceAll("\\.", "")+"\\";  
36.             String svpath = userPath+file.getFileName();  
37.             if(!file.isMissing()){  
38.                 String realPath = request.getRealPath("/");  
39.                 FileUtil.creatPath(realPath+userPath);  
40.                 file.saveAs(svpath,SmartUpload.SAVE_VIRTUAL);  
41.                 try {  
42.                     OCRUtil.runOCR(realPath, realPath+svpath, realPath+userPath+"ocr",true);  
43.                     request.setAttribute("txt", FileUtil.read(realPath+userPath+"ocr.txt").trim());  
44.                     request.getRequestDispatcher("/index.jsp").forward(request, response);  
45.                 } catch (Exception e) {  
46.                     e.printStackTrace();  
47.                 }  
48.                 FileUtil.delete(realPath+userPath);  
49.             }  
50.         } catch (SmartUploadException e) {  
51.             e.printStackTrace();  
52.         }  
53.     }  
54.   
55. }  

package servlet;

import java.io.IOException;

import javax.servlet.ServletConfig;
import javax.servlet.ServletException;
import javax.servlet.http.HttpServlet;
import javax.servlet.http.HttpServletRequest;
import javax.servlet.http.HttpServletResponse;

import util.FileUtil;
import util.OCRUtil;

import com.jspsmart.upload.File;
import com.jspsmart.upload.SmartUpload;
import com.jspsmart.upload.SmartUploadException;


public class OCRServlet extends HttpServlet {

	public void doPost(HttpServletRequest request, HttpServletResponse response)
			throws ServletException, IOException {
		response.setCharacterEncoding("gbk");
		SmartUpload upload = new SmartUpload();
		ServletConfig sc = this.getServletConfig();
		upload.initialize(sc, request, response);
		File file = null;
		long size = 5*1024*1024;
		upload.setAllowedFilesList("gif,jpg,bmp,png");
		upload.setMaxFileSize(size);
		upload.setCharset("GBK");
		try {
			upload.upload();
			file = upload.getFiles().getFile(0);
			String userPath = "upload\\"+request.getRemoteAddr().replaceAll("\\.", "")+"\\";
			String svpath = userPath+file.getFileName();
			if(!file.isMissing()){
				String realPath = request.getRealPath("/");
				FileUtil.creatPath(realPath+userPath);
				file.saveAs(svpath,SmartUpload.SAVE_VIRTUAL);
				try {
					OCRUtil.runOCR(realPath, realPath+svpath, realPath+userPath+"ocr",true);
					request.setAttribute("txt", FileUtil.read(realPath+userPath+"ocr.txt").trim());
					request.getRequestDispatcher("/index.jsp").forward(request, response);
				} catch (Exception e) {
					e.printStackTrace();
				}
				FileUtil.delete(realPath+userPath);
			}
		} catch (SmartUploadException e) {
			e.printStackTrace();
		}
	}

}

[java] view plain copy

print ?

1. package util;  
2.   
3. public class OCRUtil {  
4.     public static String chiSIM = "chi_sim";  
5.       
6.     public static void runOCR(String realPath,String imagePath,String outPath,boolean isChi) throws Exception{  
7.         Runtime r = Runtime.getRuntime();  
8.         String cmd = "\""+realPath+"Tesseract-OCR\\tesseract.exe\" \""+imagePath+"\" \""+outPath+"\" -l "+(isChi?chiSIM:"");  
9.         r.exec(cmd);  
10.     }  
11. }  

package util;

public class OCRUtil {
	public static String chiSIM = "chi_sim";
	
	public static void runOCR(String realPath,String imagePath,String outPath,boolean isChi) throws Exception{
		Runtime r = Runtime.getRuntime();
		String cmd = "\""+realPath+"Tesseract-OCR\\tesseract.exe\" \""+imagePath+"\" \""+outPath+"\" -l "+(isChi?chiSIM:"");
		r.exec(cmd);
	}
}

[java] view plain copy

print ?

1. package util;  
2.   
3. import java.io.BufferedReader;  
4. import java.io.File;  
5. import java.io.FileInputStream;  
6. import java.io.IOException;  
7. import java.io.InputStreamReader;  
8.   
9. public class FileUtil {  
10.     public static String read(String path) throws IOException{  
11.         String txt = "";  
12.         File file = new File(path);  
13.         long timeout = 30*60;  
14.         while(!(file.isFile() && file.exists())){  
15.             file = new File(path);  
16.             try {  
17.                 Thread.sleep(100);  
18.                 timeout -= 100;  
19.             } catch (InterruptedException e) {  
20.                 e.printStackTrace();  
21.             }  
22.         }  
23.         if (file.isFile() && file.exists()) {  
24.             InputStreamReader read = new InputStreamReader(new FileInputStream(file), "UTF-8");  
25.             BufferedReader bReader = new BufferedReader(read);  
26.             String temptxt = "";  
27.             txt = "";  
28.             while((temptxt=bReader.readLine())!=null){  
29.                 txt += temptxt;  
30.             }  
31.             bReader.close();  
32.             read.close();  
33.         }  
34.         return txt;  
35.     }  
36.       
37.     public static void creatPath(String path) throws IOException{  
38.         File file = new File(path);  
39.         file.mkdir();  
40.     }  
41.       
42.     public static void delete(String path) throws IOException{  
43.         File file = new File(path);  
44.         String[] list = file.list();  
45.         File tempFile = null;  
46.         for(String temp : list){  
47.             tempFile = new File(path+temp);  
48.             tempFile.delete();  
49.         }  
50.         file.delete();  
51.     }  
52. }  

package util;

import java.io.BufferedReader;
import java.io.File;
import java.io.FileInputStream;
import java.io.IOException;
import java.io.InputStreamReader;

public class FileUtil {
	public static String read(String path) throws IOException{
		String txt = "";
		File file = new File(path);
		long timeout = 30*60;
		while(!(file.isFile() && file.exists())){
			file = new File(path);
			try {
				Thread.sleep(100);
				timeout -= 100;
			} catch (InterruptedException e) {
				e.printStackTrace();
			}
		}
		if (file.isFile() && file.exists()) {
			InputStreamReader read = new InputStreamReader(new FileInputStream(file), "UTF-8");
			BufferedReader bReader = new BufferedReader(read);
			String temptxt = "";
			txt = "";
			while((temptxt=bReader.readLine())!=null){
				txt += temptxt;
			}
			bReader.close();
			read.close();
		}
		return txt;
	}
	
	public static void creatPath(String path) throws IOException{
		File file = new File(path);
		file.mkdir();
	}
	
	public static void delete(String path) throws IOException{
		File file = new File(path);
		String[] list = file.list();
		File tempFile = null;
		for(String temp : list){
			tempFile = new File(path+temp);
			tempFile.delete();
		}
		file.delete();
	}
}

下面是JSP代码：

[java] view plain copy

print ?

1. <%@ page language="java" import="java.util.*" pageEncoding="GBK"%>  
2.   
3. <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">  
4. <html>  
5.   <head>  
6.     <title>在线OCR--By Lee</title>  
7.       
8.     <meta http-equiv="pragma" content="no-cache">  
9.     <meta http-equiv="cache-control" content="no-cache">  
10.     <meta http-equiv="expires" content="0">      
11.     <meta http-equiv="keywords" content="keyword1,keyword2,keyword3">  
12.     <meta http-equiv="description" content="This is my page">  
13.     <!--a  
14.     <link rel="stylesheet" type="text/css" href="styles.css">  
15.     -->  
16.   
17.   </head>  
18.   <script type="text/javascript">  
19.     function check(){  
20.         var path = document.getElementById("image").value;  
21.         if(path.length==0){  
22.             alert("请选择要导入的图片！");  
23.             return false;  
24.         }  
25.         if(!(path.match(/.jpg$/i)||path.match(/.bmp$/i)||path.match(/.gif$/i)||path.match(/.png$/i))){  
26.             alert("只支持JPG,BMP,GIF,PNG格式！");  
27.             return false;  
28.         }  
29.         return true;  
30.     }  
31.   </script>  
32.   <body>  
33.   <form enctype="multipart/form-data" method="post" action="OCRServlet" οnsubmit="return check();">  
34.      选择文件：<input type="file" id="image" name="image"><br/>  
35.      上传文件：<input type="submit" value="提交上传">  
36.   </form>  
37.   <textarea rows="20" cols="60"><%Object txt = request.getAttribute("txt");   
38.     if(txt!=null&&txt.toString().length()==0){  
39.         out.print("未识别出任何文字！");  
40.     }else if(txt!=null){  
41.         out.print(txt.toString());  
42.     }  
43.   %></textarea>  
44.   </body>  
45. </html>  

<%@ page language="java" import="java.util.*" pageEncoding="GBK"%>

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
  <head>
    <title>在线OCR--By Lee</title>
    
	<meta http-equiv="pragma" content="no-cache">
	<meta http-equiv="cache-control" content="no-cache">
	<meta http-equiv="expires" content="0">    
	<meta http-equiv="keywords" content="keyword1,keyword2,keyword3">
	<meta http-equiv="description" content="This is my page">
	<!--a
	<link rel="stylesheet" type="text/css" href="styles.css">
	-->

  </head>
  <script type="text/javascript">
  	function check(){
  		var path = document.getElementById("image").value;
  		if(path.length==0){
  			alert("请选择要导入的图片！");
  			return false;
  		}
  		if(!(path.match(/.jpg$/i)||path.match(/.bmp$/i)||path.match(/.gif$/i)||path.match(/.png$/i))){
  			alert("只支持JPG,BMP,GIF,PNG格式！");
  			return false;
  		}
  		return true;
  	}
  </script>
  <body>
  <form enctype="multipart/form-data" method="post" action="OCRServlet" οnsubmit="return check();">
     选择文件：<input type="file" id="image" name="image"><br/>
     上传文件：<input type="submit" value="提交上传">
  </form>
  <textarea rows="20" cols="60"><%Object txt = request.getAttribute("txt"); 
  	if(txt!=null&&txt.toString().length()==0){
  		out.print("未识别出任何文字！");
  	}else if(txt!=null){
  		out.print(txt.toString());
  	}
  %></textarea>
  </body>
</html>

效果图：

在图片没做任何处理的情况下，识别率还是挺低的。。

• HTML/XML objective-c Delphi Ruby PHP C# C++ JavaScript