java实现识别图片上的文字（OCR识别身份证等证件信息）

dong__csdn

已于 2024-04-19 17:01:30 修改

阅读量3.2k

点赞数 11

文章标签： java ocr

于 2024-04-19 16:29:00 首次发布

本文链接：https://blog.csdn.net/dong__CSDN/article/details/137968865

版权

利用第三方jar包，实现识别图片上的文字。第三方支持地址：Spire.OCR for Java | 专业的图文识别组件，用以读取图片格式中的文本Spire.OCR for Java 是专为 Java 开发者设计的强大OCR库，提供高效的文字识别功能，能够从图像中准确识别和提取文本。https://www.e-iceblue.cn/Introduce/Spire-OCR-JAVA.html

1.maven项目，添加依赖

<dependency>
    <groupId>e-iceblue</groupId>
    <artifactId>spire.ocr</artifactId>
    <version>1.9.0</version>
</dependency>

<repositories>
    <repository>
        <id>com.e-iceblue</id>
        <name>e-iceblue</name>
        <url>https://repo.e-iceblue.cn/repository/maven-public/</url>
    </repository>
</repositories>

2.根据服务器类型，下载相关依赖文件

linux

windows

3.编写demo代码

    @Test
    public void getText() throws OcrException {
        //指定依赖文件的路径
        String dependencies = "dependencies/";
        //指定要需要扫描的图片的路径
        String imageFile = "C:\\Users\\user\\Desktop\\3.jpg";
        //指定输出文件的路径
        String outputFile = "test.txt";

        //创建OcrScanner对象，并设置其依赖文件路径
        OcrScanner scanner = new OcrScanner();
        scanner.setDependencies(dependencies);

        //扫描指定的图像文件
        scanner.scan(imageFile);
        //获取扫描的文本内容
        String scannedText = scanner.getText().toString();
        System.out.println(scannedText);
    }

4.扩展

根据图片OCR识别，来获取身份证等证件图片上的信息，并用正则表达式解析识别的文字

    @Test
    public void testIdcard() throws OcrException {
        String filePath = "C:\\Users\\85753\\user\\idcard\\";
        for (int i = 1; i < 11; i++) {
            long startTime = System.currentTimeMillis();
            HashMap<String, String> hashMap = idCardOCR(filePath + i + ".jpg");
            long endTime = System.currentTimeMillis();
            System.out.println(hashMap);
            System.out.println("一共耗时：" + (startTime - endTime) + "毫秒！");
        }
    }

    public HashMap<String, String> idCardOCR(String filePath) throws OcrException {
        //指定依赖文件的路径
        String dependencies = "dependencies/";
        //指定要需要扫描的图片的路径
        String imageFile = filePath;

        //创建OcrScanner对象，并设置其依赖文件路径
        OcrScanner scanner = new OcrScanner();
        scanner.setDependencies(dependencies);
        scanner.scan(imageFile);//扫描指定的图像文件

        //获取扫描的文本内容
        String scannedText = scanner.getText().toString();

        // 去除换行，去除无用数据
        String result = scannedText.replaceAll("\r", "").replaceAll("\n", "");
        result = result.replaceAll("Evaluation Warning : The version can be used only for     evaluation purpose...", "");
        System.out.println(result);

        HashMap<String, String> map = new HashMap<>();
        map = this.dealIDCardText(result);

        return map;
    }

    /**
     * 正则处理识别的身份证数据
     *
     * @param result
     * @return
     */
    public static HashMap<String, String> dealIDCardText(String result) {
        HashMap<String, String> map = new HashMap<>();
        String name = "";
        String sex = "";
        String nation = "";
        String birth = "";
        String address = "";
        String idCard = "";
        // 处理数据(正则表达式获取身份证信息，满足以下一条即可)
        String regex = "姓名(\\S+)民族(\\S+)性别(\\S+)出生(\\d{4}年\\d{1,2}月\\d{1,2}日)住址(\\S+)公民身份号码(\\d{17}[0-9Xx])";
        Pattern pattern = Pattern.compile(regex);
        Matcher matcher = pattern.matcher(result);
        String regex2 = "(\\S+)姓名民族(\\S+)性别(\\S+)出生(\\d{4}年\\d{1,2}月\\d{1,2}日)住址(\\S+)(\\d{17}[0-9Xx])公民身份号码";
        Pattern pattern2 = Pattern.compile(regex2);
        Matcher matcher2 = pattern2.matcher(result);
        String regex3 = "姓名(\\S+)性别(\\S+)民族(\\S+)出生(\\d{4}年\\d{1,2}月\\d{1,2}日)住址(\\S+)公民身份号码(\\d{17}[0-9Xx])";
        Pattern pattern3 = Pattern.compile(regex3);
        Matcher matcher3 = pattern3.matcher(result);
        String regex4 = "姓名(\\S+)性别(\\S+)民族(\\S+)出生(\\d{4}年\\d{1,2}月\\d{1,2}日)住址(\\S+)(\\d{17}[0-9Xx])公民身份号码";
        Pattern pattern4 = Pattern.compile(regex4);
        Matcher matcher4 = pattern4.matcher(result);
        String regex5 = "姓名(\\S+)民族(\\S+)性别(\\S+)出生(\\d{4}年\\d{1,2}月\\d{1,2}日)住址(\\S+)(\\d{17}[0-9Xx])公民身份号码";
        Pattern pattern5 = Pattern.compile(regex5);
        Matcher matcher5 = pattern5.matcher(result);
        String regex6 = "(\\S+)姓名性别(\\S+)民族(\\S+)出生(\\d{4}年\\d{1,2}月\\d{1,2}日)住址(\\S+)(\\d{17}[0-9Xx])公民身份号码";
        Pattern pattern6 = Pattern.compile(regex6);
        Matcher matcher6 = pattern6.matcher(result);

        if (matcher.find()) {
            name = matcher.group(1);
            nation = matcher.group(2);
            sex = matcher.group(3);
            birth = matcher.group(4);
            address = matcher.group(5);
            idCard = matcher.group(6);
        } else if (matcher2.find()) {
            name = matcher2.group(1);
            nation = matcher2.group(2);
            sex = matcher2.group(3);
            birth = matcher2.group(4);
            address = matcher2.group(5);
            idCard = matcher2.group(6);
        } else if (matcher3.find()) {
            name = matcher3.group(1);
            sex = matcher3.group(2);
            nation = matcher3.group(3);
            birth = matcher3.group(4);
            address = matcher3.group(5);
            idCard = matcher3.group(6);
        } else if (matcher4.find()) {
            name = matcher4.group(1);
            sex = matcher4.group(2);
            nation = matcher4.group(3);
            birth = matcher4.group(4);
            address = matcher4.group(5);
            idCard = matcher4.group(6);
        } else if (matcher5.find()) {
            name = matcher5.group(1);
            nation = matcher5.group(2);
            sex = matcher5.group(3);
            birth = matcher5.group(4);
            address = matcher5.group(5);
            idCard = matcher5.group(6);
        } else if (matcher6.find()) {
            name = matcher6.group(1);
            sex = matcher6.group(2);
            nation = matcher6.group(3);
            birth = matcher6.group(4);
            address = matcher6.group(5);
            idCard = matcher6.group(6);
        }

        map.put("name", name);
        map.put("sex", sex);
        map.put("nation", nation);
        map.put("birth", birth);
        map.put("address", address);
        map.put("idCard", idCard);

        return map;
    }

注意：因图片识别后的身份证数据顺序不固定，经过多次测试，将可能的结果通过正则表达式来解析。如果符合其中一个正则，就将匹配的数据返回。

参考：如何在 Java 项目中扫描识别图片中的文字