参考:Tess4J 简单使用入门_一个爱浪费时间的人的博客-CSDN博客_tess4j
一.简介
Tesseract-OCR支持中文识别,并且开源和提供全套的训练工具,是快速低成本开发的首选。而Tess4J则是Tesseract在Java PC上的应用。在英文和数字识别中性能还是不错的,但是在中文识别中,无论速度还是识别率还是较弱,建议有条件的话,针对场景进行训练,会获得较好结果,本文仅对目前Tess4J的用法进行介绍
二.入门教程
2.1 Tess4J官网下载最新的源码包
Github地址:GitHub - nguyenq/tess4j: Java JNA wrapper for Tesseract OCR API
2.2 Tess4J目录说明
- Tess4J目录:
- dist:综合jar,不需要配置dll
- lib:所有相关的jar包
- src:源码包
- tessdata:训练好的字体模型
- test:测试用例
2.3 Java Maven Demo
Pom.xml
<?xml version="1.0" encoding="UTF-8"?>
<project xmlns="http://maven.apache.org/POM/4.0.0"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 http://maven.apache.org/xsd/maven-4.0.0.xsd">
<modelVersion>4.0.0</modelVersion>
<groupId>org.example</groupId>
<artifactId>tess4j-demo</artifactId>
<version>1.0-SNAPSHOT</version>
<properties>
<maven.compiler.source>8</maven.compiler.source>
<maven.compiler.target>8</maven.compiler.target>
<project.build.sourceEncoding>UTF-8</project.build.sourceEncoding>
</properties>
<dependencies>
<dependency>
<groupId>net.sourceforge.tess4j</groupId>
<artifactId>tess4j</artifactId>
<version>5.4.0</version>
</dependency>
<dependency>
<groupId>junit</groupId>
<artifactId>junit-dep</artifactId>
<version>4.11</version>
</dependency>
<dependency>
<groupId>org.slf4j</groupId>
<artifactId>slf4j-api</artifactId>
<version>2.0.3</version>
</dependency>
</dependencies>
</project>
示例1
设置引擎模式,设置识别语种,设置图片分割模式
package org.example;
import net.sourceforge.tess4j.ITesseract;
import net.sourceforge.tess4j.Tesseract;
import net.sourceforge.tess4j.TesseractException;
import java.io.File;
import java.util.ArrayList;
import java.util.List;
public class Main {
public static void main(String[] args) throws TesseractException {
System.out.println("Hello world!");
File file = new File("D:\\1.jpg");
ITesseract iTesseract = new Tesseract();
iTesseract.setDatapath("D:\\programs\\TessractOCR\\tessdata"); // 设置tessdata目录
iTesseract.setLanguage("eng"); // 指定需要识别的语种
// 只提取数字
List<String> configList = new ArrayList<>();
configList.add("digits");
iTesseract.setConfigs(configList);
iTesseract.setOcrEngineMode(1); // 设置OCR引擎模式(OEM)
iTesseract.setPageSegMode(6); // 设置图片分割模式(PSM)
String result = iTesseract.doOCR(file);
System.out.println(result);
}
}
- 其中模型数据就在tessdata目录下
- tessdata下载地址:traineddata,tessdata_best
示例2
对待识别图片进行预处理,便于识别
package org.example;
import net.sourceforge.tess4j.ITesseract;
import net.sourceforge.tess4j.Tesseract;
import net.sourceforge.tess4j.TesseractException;
import net.sourceforge.tess4j.util.ImageHelper;
import javax.imageio.ImageIO;
import java.awt.image.BufferedImage;
import java.io.File;
import java.io.IOException;
/**
* 图片优化
* @author 86133 2022-11-18 10:01:09
*/
public class ImgOptimize {
public static void main(String[] args) throws IOException {
// 识别图片的文件(修改为自己的图片路径)
String imagePath = ClassLoader.getSystemResource("tupian/9188.png").getPath();
if (imagePath.length() > 0) {
imagePath = imagePath.substring(1);
}
System.out.println("imagePath:" + imagePath);
File file = new File(imagePath);
// 图片转图片流
BufferedImage img = ImageIO.read(file);
// 这里对图片黑白处理,增强识别率.这里先通过截图,截取图片中需要识别的部分
img = ImageHelper.convertImageToGrayscale(img);
// 图片锐化,自己使用中影响识别率的主要因素是针式打印机字迹不连贯,所以锐化反而降低识别率
img = ImageHelper.convertImageToBinary(img);
// 图片放大5倍,增强识别率(很多图片本身无法识别,放大7倍时就可以轻易识,但是考滤到客户电脑配置低,针式打印机打印不连贯的问题,这里就放大7倍)
img = ImageHelper.getScaledInstance(img, img.getWidth() * 7, img.getHeight() * 7);
ITesseract instance = new Tesseract();
//设置训练库的位置
String path = ClassLoader.getSystemResource("tessdata").getPath();
if (path.length() > 0) {
path = path.substring(1);
}
// 打印路径
System.out.println("tessdata:" + path);
instance.setDatapath(path);
//chi_sim :简体中文, eng 根据需求选择语言库
instance.setLanguage("eng");
String result = null;
try {
long startTime = System.currentTimeMillis();
result = instance.doOCR(img);
long endTime = System.currentTimeMillis();
System.out.println("Time is:" + (endTime - startTime) + " 毫秒");
} catch (TesseractException e) {
e.printStackTrace();
}
System.out.println("result: " + result);
}
}
说明:
// 基于图片识别的识别率不高,一般先做图片的处理再进行识别。Tess4J专门提供的ImageHelper:
getScaledInstance 放大图片
getSubImage 截取图片
convertImageToBinary 转二进制
convertImageToGrayscale 将图像转换为灰度
invertImageColor 反转图像颜色
rotateImage 旋转影像
//图片转图片流
BufferedImage img = ImageIO.read(file);
// 这里对图片黑白处理,增强识别率.这里先通过截图,截取图片中需要识别的部分
img = ImageHelper.convertImageToGrayscale(img);
// 图片锐化,自己使用中影响识别率的主要因素是针式打印机字迹不连贯,所以锐化反而降低识别率
img = ImageHelper.convertImageToBinary(img);
// 图片放大5倍,增强识别率(很多图片本身无法识别,放大7倍时就可以轻易识,
// 但是考滤到客户电脑配置低,针式打印机打印不连贯的问题,这里就放大7倍)
img = ImageHelper.getScaledInstance(img, img.getWidth() * 7, img.getHeight() * 7);
示例3
将识别后的文本,标识到图片中相应的识别位置上
package org.example;
import net.sourceforge.tess4j.*;
import javax.imageio.ImageIO;
import javax.imageio.stream.ImageOutputStream;
import java.awt.*;
import java.awt.image.BufferedImage;
import java.io.File;
import java.io.IOException;
import java.nio.file.Files;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.List;
/**
* 图片文字识别,并将识别结果标识在图片上
*
* @author 86133 2022-11-17 15:04:18
*/
public class ImgIdentification {
public static void main(String[] args) throws IOException {
File imageFile = new File("img/test.png");
BufferedImage bi = ImageIO.read(Files.newInputStream(imageFile.toPath()));
ITesseract instance = getInstance();
// 获取识别的词汇列表
List<Word> wordList = instance.getWords(bi, ITessAPI.TessPageIteratorLevel.RIL_WORD);
System.out.println("wordList: ----------------- " + wordList);
File outImageFile = new File("img/result.png");
identifyImg(bi, wordList, outImageFile);
}
/**
* 识别并标识图片
*
* @param wordList
* @throws IOException
*/
private static void identifyImg(BufferedImage bi, List<Word> wordList, File outImageFile) throws IOException {
// 创建画笔
Graphics2D pen = bi.createGraphics();
pen.setColor(Color.RED);
for (Word word : wordList) {
Rectangle rect = word.getBoundingBox();
pen.drawRect(rect.x, rect.y, rect.width, rect.height);
pen.setFont(new Font("微软雅黑", Font.ITALIC, 18));
pen.drawString(word.getText(), rect.x, rect.y - 10);
}
ImageOutputStream ios = ImageIO.createImageOutputStream(outImageFile);
ImageIO.write(bi, "png", ios);
}
private static void createDocumentsWithResultsTest() throws Exception {
ITesseract instance = getInstance();
File imageFile = new File("img", "image.png");
BufferedImage bi = ImageIO.read(Files.newInputStream(imageFile.toPath()));
String outputbase = "target/test-classes/test-results/docrenderer1-5";
List<ITesseract.RenderedFormat> formats = new ArrayList<>(Arrays.asList(ITesseract.RenderedFormat.HOCR, ITesseract.RenderedFormat.PDF, ITesseract.RenderedFormat.TEXT));
OCRResult or = instance.createDocumentsWithResults(bi, imageFile.getPath(), outputbase, formats, ITessAPI.TessPageIteratorLevel.RIL_WORD);
System.out.println(or.getWords());
}
private static ITesseract getInstance() {
ITesseract instance = new Tesseract();
// 设置tessdata目录, 下载地址: https://github.com/tesseract-ocr/tessdata_best
instance.setDatapath("D:\\programs\\TessractOCR\\tessdata");
/**
* 设置需要识别的语种,
* chi_sim+eng: 表示识别简体中文和英文;
* eng: 仅识别英文
*/
instance.setLanguage("chi_sim+eng");
instance.setOcrEngineMode(1); // 设置OCR引擎模式(OEM)
instance.setPageSegMode(6); // 设置图片分割模式(PSM)
return instance;
}
}
2.4 READ.ME
# Note
Tesseract引入训练模型的方法
根据自己的需要下载需要的模型文件,将traineddata文件放在
%TesseractOCR_HOME%\tessdata 目录(Tesseract安装目录)下就可以了。
tessdata下载地址:
https://github.com/tesseract-ocr/tessdata_best
tessdata_best可用来再训练字库
训练方法参考文档:
https://tesseract-ocr.github.io/tessdoc/tess4/TrainingTesseract-4.00.html
# 参数释义
## 自定义配置
config = r'-l chi_sim+eng --psm 6'
# 其它OCR选项:
# --tessdata-dir PATH:Specify the location of tessdata path.
# --user-words PATH:Specify the location of user words file.
# --user-patterns PATH:Specify the location of user patterns file.
# --dpi VALUE:Specify DPI for input image.
# -l LANG[+LANG]:Specify language(s) used for OCR.
# -c VAR=VALUE:Set value for config variables. Multiple -c arguments are allowed.
# --psm NUM:Specify page segmentation mode.
# --oem NUM:Specify OCR Engine mode.
text = pytesseract.image_to_string(Image.open(r'D:\workspace\tesseract-ocr\test.jpg'), config=config)
## ImageHelper
getScaledInstance 放大图片
getSubImage 截取图片
convertImageToBinary 转二进制
convertImageToGrayscale 将图像转换为灰度
invertImageColor 反转图像颜色
rotateImage 旋转影像
## 图片分割模式(PSM)
tesseract有13种图片分割模式(page segmentation mode,psm):
0 – Orientation and script detection (OSD) only. 方向及语言检测(Orientation and script detection,OSD)
1 – Automatic page segmentation with OSD. 自动图片分割
2 – Automatic page segmentation, but no OSD, or OCR. 自动图片分割,没有OSD和OCR
3 – Fully automatic page segmentation, but no OSD. (Default) 完全的自动图片分割,没有OSD
4 – Assume a single column of text of variable sizes. 假设有一列不同大小的文本
5 – Assume a single uniform block of vertically aligned text. 假设有一个垂直对齐的文本块
6 – Assume a single uniform block of text. 假设有一个对齐的文本块
7 – Treat the image as a single text line. 图片为单行文本
8 – Treat the image as a single word. 图片为单词
9 – Treat the image as a single word in a circle. 图片为圆形的单词
10 – Treat the image as a single character. 图片为单个字符
11 – Sparse text. Find as much text as possible in no particular order. 稀疏文本。查找尽可能多的文本,没有特定的顺序。
12 – Sparse text with OSD. OSD稀疏文本
13 – Raw line. Treat the image as a single text line, bypassing hacks that are Tesseract-specific. 原始行。将图像视为单个文本行。
## OCR引擎模式(OEM)
有4种OCR引擎模式:
0 – Legacy engine only.
1 – Neural nets LSTM engine only.
2 – Legacy + LSTM engines.
3 – Default, based on what is available.
## 方向及语言检测OSD
Tesseract支持方向及语言检测(Orientation and script detection,OSD) ,比如检测下面的图片:

### 检测方法:
osd = pytesseract.image_to_osd('osd-example.png',config='--psm 0 -c min_characters_to_try=5')
print(osd)
### 执行结果:
Page number: 0
Orientation in degrees: 90
Rotate: 270
Orientation confidence: 0.74
Script: Han
Script confidence: 0.83