Java tesseract-ocr 图文识别技术 Java代码实现

Java文字识别程序的关键是寻找一个可以调用

import java.awt.image.BufferedImage;
import java.io.File;
import java.io.IOException;
import java.util.Iterator;
import java.util.Locale;

import javax.imageio.IIOImage;
import javax.imageio.ImageIO;
import javax.imageio.ImageReader;
import javax.imageio.ImageWriteParam;
import javax.imageio.ImageWriter;
import javax.imageio.metadata.IIOMetadata;
import javax.imageio.stream.ImageInputStream;
import javax.imageio.stream.ImageOutputStream;

import com.sun.media.imageio.plugins.tiff.TIFFImageWriteParam;

public class ImageIOHelper {

	public static File createImage(File imageFile, String imageFormat) {
		File tempFile = null;
		try {
			Iterator readers = ImageIO.getImageReadersByFormatName(imageFormat);
			ImageReader reader = (ImageReader) readers.next();

			ImageInputStream iis = ImageIO.createImageInputStream(imageFile);
			reader.setInput(iis);
			// Read the stream metadata
			IIOMetadata streamMetadata = reader.getStreamMetadata();

			// Set up the writeParam
			TIFFImageWriteParam tiffWriteParam = new TIFFImageWriteParam(Locale.CHINESE);
			tiffWriteParam.setCompressionMode(ImageWriteParam.MODE_DISABLED);

			// Get tif writer and set output to file
			Iterator writers = ImageIO.getImageWritersByFormatName("tiff");
			ImageWriter writer = (ImageWriter) writers.next();

			BufferedImage bi = reader.read(0);
			IIOImage image = new IIOImage(bi, null, reader.getImageMetadata(0));
			tempFile = tempImageFile(imageFile);
			ImageOutputStream ios = ImageIO.createImageOutputStream(tempFile);
			writer.setOutput(ios);
			writer.write(streamMetadata, image, tiffWriteParam);
			ios.close();

			writer.dispose();
			reader.dispose();

		} catch (IOException e) {
			e.printStackTrace();
		}
		return tempFile;
	}

	private static File tempImageFile(File imageFile) {
		String path = imageFile.getPath();
		StringBuffer strB = new StringBuffer(path);
		strB.insert(path.lastIndexOf('.'), 0);
		return new File(strB.toString().replaceFirst("(?<=//.)(//w+)$", "tif"));
	}

}

的OCR引擎。tesseract-ocr就是一个这样的OCR引擎，在1985年到1995年由HP实验室开发，现在在Google。tesseract-ocr 3.0发布，支持中文。不过tesseract-ocr 3.0不是图形化界面的客户端，别人写的FreeOCR图形化客户端还不支持导入新的 3.0 traineddata。但这标志着，现在有自由的中文OCR软件了。

 

    java中使用tesseract-ocr3.01的步骤如下：

1.下载安装tesseract-ocr-setup-3.01-1.exe（3.0以上版本才增加了中文识别）

2.在安装向导中可以选择需要下载的语言包。

3.到网上搜索下载java图形处理所需的2个包：jai_imageio-1.1-alpha.jar，swingx-1.6.1.jar

4.java程序清单：

ImageIOHelper 类：

[html] view plain copy

1. import java.awt.image.BufferedImage;  
2. import java.io.File;  
3. import java.io.IOException;  
4. import java.util.Iterator;  
5. import java.util.Locale;  
6.   
7. import javax.imageio.IIOImage;  
8. import javax.imageio.ImageIO;  
9. import javax.imageio.ImageReader;  
10. import javax.imageio.ImageWriteParam;  
11. import javax.imageio.ImageWriter;  
12. import javax.imageio.metadata.IIOMetadata;  
13. import javax.imageio.stream.ImageInputStream;  
14. import javax.imageio.stream.ImageOutputStream;  
15.   
16. import com.sun.media.imageio.plugins.tiff.TIFFImageWriteParam;  
17.   
18. public class ImageIOHelper {    
19.         
20.     public static File createImage(File imageFile, String imageFormat) {    
21.         File tempFile = null;    
22.         try {    
23.             Iterator readers = ImageIO.getImageReadersByFormatName(imageFormat);    
24.             ImageReader reader = readers.next();    
25.             
26.             ImageInputStream iis = ImageIO.createImageInputStream(imageFile);    
27.             reader.setInput(iis);    
28.             //Read the stream metadata    
29.             IIOMetadata streamMetadata = reader.getStreamMetadata();    
30.                 
31.             //Set up the writeParam    
32.             TIFFImageWriteParam tiffWriteParam = new TIFFImageWriteParam(Locale.CHINESE);    
33.             tiffWriteParam.setCompressionMode(ImageWriteParam.MODE_DISABLED);    
34.                 
35.             //Get tif writer and set output to file    
36.             Iterator writers = ImageIO.getImageWritersByFormatName("tiff");    
37.             ImageWriter writer = writers.next();    
38.                 
39.             BufferedImage bi = reader.read(0);    
40.             IIOImage image = new IIOImage(bi,null,reader.getImageMetadata(0));    
41.             tempFile = tempImageFile(imageFile);    
42.             ImageOutputStream ios = ImageIO.createImageOutputStream(tempFile);    
43.             writer.setOutput(ios);    
44.             writer.write(streamMetadata, image, tiffWriteParam);    
45.             ios.close();    
46.                 
47.             writer.dispose();    
48.             reader.dispose();    
49.                 
50.         } catch (IOException e) {    
51.             e.printStackTrace();    
52.         }    
53.         return tempFile;    
54.     }    
55.     
56.     private static File tempImageFile(File imageFile) {    
57.         String path = imageFile.getPath();    
58.         StringBuffer strB = new StringBuffer(path);    
59.         strB.insert(path.lastIndexOf('.'),0);    
60.         return new File(strB.toString().replaceFirst("(?<=//.)(//w+)$", "tif"));    
61.     }    
62.     
63. }  

 

OCR 类：

[html] view plain copy

1. package com.hhp.util;  
2.   
3. import java.io.BufferedReader;    
4. import java.io.File;    
5. import java.io.FileInputStream;    
6. import java.io.InputStreamReader;    
7. import java.util.ArrayList;    
8. import java.util.List;    
9. import org.jdesktop.swingx.util.OS;    
10.     
11. public class OCR {    
12.     private final String LANG_OPTION = "-l";  //英文字母小写l，并非数字1    
13.     private final String EOL = System.getProperty("line.separator");    
14.     private String tessPath = "C://Program Files (x86)//Tesseract-OCR";    
15.     //private String tessPath = new File("tesseract").getAbsolutePath();    
16.         
17.     public String recognizeText(File imageFile,String imageFormat)throws Exception{    
18.         File tempImage = ImageIOHelper.createImage(imageFile,imageFormat);    
19.         File outputFile = new File(imageFile.getParentFile(),"output");    
20.         StringBuffer strB = new StringBuffer();    
21.         List cmd = new ArrayList();    
22.         if(OS.isWindowsXP()){    
23.             cmd.add(tessPath+"//tesseract");    
24.         }else if(OS.isLinux()){    
25.             cmd.add("tesseract");    
26.         }else{    
27.             cmd.add(tessPath+"//tesseract");    
28.         }    
29.         cmd.add("");    
30.         cmd.add(outputFile.getName());    
31.         cmd.add(LANG_OPTION);    
32.         cmd.add("chi_sim");    
33.         //cmd.add("eng");    
34.             
35.         ProcessBuilder pb = new ProcessBuilder();    
36.         pb.directory(imageFile.getParentFile());    
37.             
38.         cmd.set(1, tempImage.getName());    
39.         pb.command(cmd);    
40.         pb.redirectErrorStream(true);    
41.             
42.         Process process = pb.start();    
43.         //tesseract.exe 1.jpg 1 -l chi_sim    
44.         int w = process.waitFor();    
45.             
46.         //删除临时正在工作文件    
47.         tempImage.delete();    
48.             
49.         if(w==0){    
50.             BufferedReader in = new BufferedReader(new InputStreamReader(new FileInputStream(outputFile.getAbsolutePath()+".txt"),"UTF-8"));    
51.                 
52.             String str;    
53.             while((str = in.readLine())!=null){    
54.                 strB.append(str).append(EOL);    
55.             }    
56.             in.close();    
57.         }else{    
58.             String msg;    
59.             switch(w){    
60.                 case 1:    
61.                     msg = "Errors accessing files.There may be spaces in your image's filename.";    
62.                     break;    
63.                 case 29:    
64.                     msg = "Cannot recongnize the image or its selected region.";    
65.                     break;    
66.                 case 31:    
67.                     msg = "Unsupported image format.";    
68.                     break;    
69.                 default:    
70.                     msg = "Errors occurred.";    
71.             }    
72.             tempImage.delete();    
73.             throw new RuntimeException(msg);    
74.         }    
75.         new File(outputFile.getAbsolutePath()+".txt").delete();    
76.         return strB.toString();    
77.     }    
78. }    
79.   
80.   
81.    

测试类TestOCR ：

 

[html] view plain copy

1. import java.io.File;  
2. import java.io.IOException;  
3.   
4. import com.hhp.util.OCR;  
5.   
6. public class OcrTest {  
7.   
8.  public static void main(String[] args) {  
9.         String path = "C://temp//OCRcode//4.png";       
10.         System.out.println("ORC Test Begin......");  
11.         try {       
12.             String valCode = new OCR().recognizeText(new File(path), "png");       
13.             System.out.println(valCode);       
14.         } catch (IOException e) {       
15.             e.printStackTrace();       
16.         } catch (Exception e) {    
17.             e.printStackTrace();    
18.         }         
19.         System.out.println("ORC Test End......");  
20.     }    
21.   
22. }  

经过测试，tesseract-ocr 3.01的文字识别率很高，对于网站中常见的验证码识别率也很高。