tess4j OCR文字识别配合HanLP实现分词、关键字提取

HanLP: Han Language Processing

地址：https://github.com/hankcs/HanLP/tree/1.x

tess4j

依赖

<!-- https://mvnrepository.com/artifact/net.sourceforge.tess4j/tess4j -->
<dependency>
    <groupId>net.sourceforge.tess4j</groupId>
    <artifactId>tess4j</artifactId>
    <version>4.4.0</version>
</dependency>

语言库下载

https://github.com/tesseract-ocr/tessdata

Demo

package com.example.demo;

import com.alibaba.fastjson.JSON;
import com.fasterxml.jackson.databind.ObjectMapper;
import com.hankcs.hanlp.HanLP;
import com.hankcs.hanlp.tokenizer.IndexTokenizer;
import com.hankcs.hanlp.tokenizer.NLPTokenizer;
import net.sourceforge.tess4j.Tesseract;
import org.springframework.web.bind.annotation.GetMapping;
import org.springframework.web.bind.annotation.RestController;
import net.sourceforge.tess4j.ITesseract;
import java.io.File;
import java.util.*;

/**
 * @author 贾镇泽 [email protected]
 * @packagename com.example.demo
 * @date 2020/11/4
 * @time 8:57
 */
@RestController
public class FileCheck {
    @GetMapping("/test")
    public String test() throws Exception
    {
        /*源文件地址*/
        File file = new File("D:\\d7e2d8a5-8322-4237-b804-33ce4a33fdda.png");
        /*实例化Tess4j*/
        ITesseract instance = new Tesseract();
        /*设置tessdata 学习库地址*/
        instance.setDatapath("E:\\Code_tmp\\tessdata");//设置你的Tess4J下的tessdata目录
        /*识别语种*/
        instance.setLanguage("chi_sim");//指定需要识别的语种
        long l = System.currentTimeMillis();
        /*返回识别内容*/
        String result = instance.doOCR(file);
        result = result.replace(" ","");
        System.out.println(HanLP.segment(result));
        System.out.println(result);
        System.out.println("用时："+(System.currentTimeMillis()-l) + "ms");
        Map<String, Object> fh = new HashMap<String, Object>();
        /*HanLp基本用法*/
        fh.put("标准分词",HanLP.segment(result));
//        fh.put("NLP分词", NLPTokenizer.segment(result));
      //  fh.put("NLP分词1", NLPTokenizer.analyze(result).toSimpleWordList());
      //  fh.put("NLP分词2", NLPTokenizer.analyze(result));
        fh.put("索引分词", IndexTokenizer.segment(result));
        fh.put("关键词提取",HanLP.extractKeyword(result,10));
        fh.put("自动摘要",HanLP.extractSummary(result,10));
        fh.put("短语提取",HanLP.extractPhrase(result,10));

        return JSON.toJSONString(fh);
    }

}

Tess4j 识别文字识别率还算可以识别表格惨不忍睹。。