feat: 新增tesseract OCR服务接口
This commit is contained in:
@@ -0,0 +1,11 @@
|
||||
package com.klp.service;
|
||||
|
||||
public interface ITesseractOcrService {
|
||||
|
||||
/**
|
||||
* 识别网络图片
|
||||
* @param imageUrl 图片URL
|
||||
* @return 识别结果
|
||||
*/
|
||||
String recognizeText(String imageUrl) throws Exception;
|
||||
}
|
||||
@@ -0,0 +1,112 @@
|
||||
package com.klp.service.impl;
|
||||
|
||||
import com.klp.common.config.TesseractConfig;
|
||||
import com.klp.service.ITesseractOcrService;
|
||||
import net.sourceforge.tess4j.ITesseract;
|
||||
import net.sourceforge.tess4j.Tesseract;
|
||||
import org.springframework.stereotype.Service;
|
||||
|
||||
import javax.annotation.Resource;
|
||||
import javax.imageio.ImageIO;
|
||||
import java.awt.*;
|
||||
import java.awt.image.BufferedImage;
|
||||
import java.io.InputStream;
|
||||
import java.net.URL;
|
||||
|
||||
@Service
|
||||
public class TesseractOcrServiceImpl implements ITesseractOcrService {
|
||||
|
||||
@Resource
|
||||
private TesseractConfig tesseractConfig;
|
||||
@Override
|
||||
public String recognizeText(String imageUrl) throws Exception {
|
||||
// 读取网络图片为 BufferedImage
|
||||
URL url = new URL(imageUrl);
|
||||
InputStream inputStream = url.openStream();
|
||||
BufferedImage image = ImageIO.read(inputStream);
|
||||
inputStream.close();
|
||||
// 预处理图片
|
||||
BufferedImage bufferedImage = preprocessImage(image);
|
||||
System.out.println("开始OCR识别...");
|
||||
// 创建Tesseract对象
|
||||
ITesseract tesseract = new Tesseract();
|
||||
// 设置字体库路径(绝对路径)
|
||||
tesseract.setDatapath(tesseractConfig.getDatapath());
|
||||
// 设置语言简体中文
|
||||
tesseract.setLanguage("chi_sim");
|
||||
// 优化OCR配置
|
||||
tesseract.setPageSegMode(6); // 假设统一的文本块
|
||||
tesseract.setOcrEngineMode(3); // 使用默认引擎
|
||||
// 执行OCR识别图片
|
||||
String result = tesseract.doOCR(bufferedImage);
|
||||
|
||||
// 清理和格式化结果
|
||||
String cleanedResult = cleanOcrResult(result);
|
||||
|
||||
System.out.println("\n=== 识别结果 ===");
|
||||
System.out.println(cleanedResult);
|
||||
|
||||
return cleanedResult;
|
||||
}
|
||||
|
||||
/**
|
||||
* 灰度化图像(直接处理 BufferedImage)
|
||||
* @param image 原始图片
|
||||
* @return 灰度图
|
||||
*/
|
||||
private static BufferedImage preprocessImage(BufferedImage image) {
|
||||
BufferedImage grayImage = new BufferedImage(
|
||||
image.getWidth(), image.getHeight(), BufferedImage.TYPE_BYTE_GRAY);
|
||||
|
||||
Graphics2D g2d = grayImage.createGraphics();
|
||||
g2d.drawImage(image, 0, 0, null);
|
||||
g2d.dispose();
|
||||
|
||||
return grayImage;
|
||||
}
|
||||
|
||||
/**
|
||||
* 清理OCR识别结果
|
||||
* @param result 原始识别结果
|
||||
* @return 清理后的结果
|
||||
*/
|
||||
private static String cleanOcrResult(String result) {
|
||||
if (result == null || result.trim().isEmpty()) {
|
||||
return "";
|
||||
}
|
||||
|
||||
// 替换常见的OCR错误
|
||||
String cleaned = result
|
||||
// 修复常见的OCR错误
|
||||
.replaceAll("英声租", "产品名称")
|
||||
.replaceAll("库咤埕号", "产品型号")
|
||||
.replaceAll("产晓序列号", "产品序列号")
|
||||
.replaceAll("购买纳证缉号", "购买凭证编号")
|
||||
.replaceAll("质保条孰", "质保条款")
|
||||
.replaceAll("本亢命", "本产品")
|
||||
.replaceAll("质僚服务", "质保服务")
|
||||
.replaceAll("质保朝内", "质保期内")
|
||||
.replaceAll("团素", "因素")
|
||||
.replaceAll("质量闰题", "质量问题")
|
||||
.replaceAll("免贵维修", "免费维修")
|
||||
.replaceAll("更挺服", "更换服")
|
||||
.replaceAll("不包挂", "不包括")
|
||||
.replaceAll("溢用", "滥用")
|
||||
.replaceAll("探环", "损坏")
|
||||
.replaceAll("取葛保管", "妥善保管")
|
||||
.replaceAll("雷凭吊证明", "需凭此证明")
|
||||
.replaceAll("客户信恩", "客户信息")
|
||||
.replaceAll("姆钗", "姓名")
|
||||
.replaceAll("联系电语", "联系电话")
|
||||
.replaceAll("电子邹件", "电子邮件")
|
||||
.replaceAll("地抛", "地址")
|
||||
.replaceAll("摒权代表", "授权代表")
|
||||
.replaceAll("介为示例", "仅为示例")
|
||||
.replaceAll("根揪实际情况", "根据实际情况")
|
||||
.replaceAll("调教", "调整")
|
||||
// 移除多余的空格
|
||||
.trim();
|
||||
|
||||
return cleaned;
|
||||
}
|
||||
}
|
||||
@@ -6,7 +6,6 @@ import com.klp.common.core.domain.PageQuery;
|
||||
import com.baomidou.mybatisplus.extension.plugins.pagination.Page;
|
||||
import com.baomidou.mybatisplus.core.conditions.query.LambdaQueryWrapper;
|
||||
import com.baomidou.mybatisplus.core.conditions.query.QueryWrapper;
|
||||
import com.baomidou.mybatisplus.core.toolkit.Wrappers;
|
||||
import com.klp.common.utils.StringUtils;
|
||||
import lombok.RequiredArgsConstructor;
|
||||
import org.springframework.stereotype.Service;
|
||||
|
||||
Reference in New Issue
Block a user